The two-token separation rule is the design constraint that when a project issues two tokens, the demand for one must never be load-bearing for the other. The test is a thought experiment with a single question: if the second token went to zero tomorrow, does the first one still do its job. Independence is the entire reason to run two tokens instead of one, and a design that couples them has paid the complexity cost of a split without buying the benefit.
Two tokens do not create a sink, a demand source or a safety margin. They create two supply schedules, two holder bases and one more place for a dependency to hide.
What separation is supposed to buy
A split makes sense when two functions have genuinely different requirements. A unit that must hold a stable value cannot also be the speculative claim on the venture's success, because those are opposite jobs. MakerDAO's whitepaper documents the canonical version: DAI as the stablecoin and MKR as the governance and recapitalisation token, with different roles and different holders.1
That is the benefit. The cost is real and usually understated. You now maintain two supply schedules, two liquidity surfaces, two sets of holder expectations and two governance conversations. If the split does not remove a specific failure mode, you have bought all of that for nothing.
A split does not fix a demand problem, and can hide one
Axie Infinity is the most studied game-economy example, with AXS as the governance token and SLP as the in-game reward earned and spent in breeding.2 Naavik's analysis of the model sets out the sustainability problem directly: emission of the reward token was not tied to a durable sink, so the split organised the economy without solving what the economy needed to absorb.3
That is the pattern worth internalising. Moving emissions onto a second token changes which chart goes down. It does not create anyone who has to buy. If your reason for splitting is that the main token's emissions look heavy, the split is cosmetic and the arithmetic follows you into the new token.
Where coupling sneaks back in
The obvious violation is mechanical: the stable unit's redemption depends on minting the other token, so a fall in one becomes a supply shock in the other and the two collapse together. That failure has documented academic postmortems and it has its own page under seigniorage, so we will not relitigate it here.
The quieter violations are the ones that survive review. The platform token must be held to claim yield generated by the asset token. The asset token's backing sits in a treasury denominated in the platform token. Marketing describes the platform token's value as derived from growth in the asset token. None of these is a redemption mechanism, and all three reintroduce correlation. Coupling is a property of the balance sheet and the narrative, not just of the contract.
How we run the test
Set the second token's price to zero on paper and walk the system forward. Does the stable unit still redeem. Can the protocol still pay for security. Does the collateral still cover the claims. Is anyone still required to buy anything. Where the answer to all of these is yes, the separation holds and the split is doing work.
Where the answer changes, name the channel. It will be one of three: a redemption path, a treasury holding, or an incentive that only exists while the second token has value. Each has a different fix, and none of them is fixed by writing the word independent in the whitepaper.
Common questions
What is the two-token separation rule?
It is the design constraint that two tokens issued by the same project must not depend on each other. The test is whether the first token still does its job if the second goes to zero. The rule exists because independence is the only reason to accept the cost of a split. Coupled tokens reintroduce the single point of failure the split was meant to remove.
Should my project use two tokens?
Only if two functions have genuinely incompatible requirements, such as a unit that must hold stable value alongside a claim on the venture's upside. MakerDAO splits a stablecoin from a governance and recapitalisation token for that reason.1 Splitting to make an emission schedule look lighter is cosmetic. The emissions move to the new token and the demand problem moves with them.
Why did dual-token game economies struggle?
Because the split organised the economy without creating demand. Axie Infinity ran governance and reward tokens separately, and Naavik's analysis points to the reward token being emitted without a durable sink to absorb it.23 A second token gives you somewhere to put emissions. It does not give you anyone who has to buy them, and that is the constraint that binds.
See Tokenomics Design Services for how this applies in practice.
Sources
- The Maker Protocol White Paper
MakerDAO, 2020
Documents the DAI stablecoin and MKR governance and recapitalisation split, and the stated rationale for separating the two roles. - Axie Infinity Whitepaper
Sky Mavis
Primary documentation of the AXS governance token and SLP in-game reward token split. - Axie Infinity: Infinite Opportunity or Infinite Peril?
Naavik
Game-economy specialist analysis of the dual-token model, including the sustainability problem created when the reward token's emission is not tied to a durable sink.
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.