Skip to content

How Futarchy Works

This page explains the core idea behind futarchy and how Futarchy Labs is exploring it.

It is recommended reading before diving into governance usage or trading details.

The Problem Futarchy Tries to Solve

Important decisions are hard.

Proposals often involve complex, uncertain tradeoffs, and it is difficult to know in advance which actions will actually improve outcomes. Votes, committees, and expert reviews can express preferences, but they do not always aggregate information about consequences very well.

Futarchy approaches this problem by asking a different question:

Instead of asking people how they want to vote, what if we ask them what they think will create value?

A useful shorthand: a normal prediction market is about betting on what will happen; futarchy is about betting on what should happen, relative to a chosen objective. More precisely, traders price which decision they expect to better achieve the goal the system has chosen.

Values vs. Beliefs

Futarchy separates two roles that are often mixed together:

  • Values: What outcomes matter
  • Beliefs: Which actions will best achieve those outcomes

Humans decide values. Markets aggregate beliefs.

For a for-profit company, the objective might be share price, enterprise value, revenue, or another measure of long-term business success. For a DAO or crypto organization, the closest market-priced analogue is often token value. For a nonprofit, the objective might be marginal impact per dollar. For an agent team, it might be task success, reliability, security, or issue-resolution rate.

The deeper pattern is that a system chooses an objective, then markets estimate which action is most likely to improve it. The important part is that the objective and measurement process are defined outside the market being used to choose actions.

This is why futarchy can apply beyond token governance. In a public-goods or nonprofit setting, for example, donors could define an impact metric and markets could estimate which proposal produces the most marginal impact per dollar. See Objectives and Impact Markets.

The general pattern

A futarchic mechanism has five basic parts:

  1. A community, company, protocol, or agent team defines what it wants to improve.
  2. A proposal or action set is put forward.
  3. Markets are opened on the possible worlds created by those actions.
  4. Traders price which world is expected to better achieve the chosen objective.
  5. The system uses the market signal to guide a decision.

That decision might remain advisory, or it might control funding, execution, treasury routing, task assignment, or other incentives.

Counterfactual Worlds

For each decision, futarchy compares possible worlds.

In a simple proposal market, the two worlds are:

  • YES - the proposal is approved
  • NO - the proposal is not approved

Other markets could compare product launches, hiring plans, grant proposals, agent configurations, model choices, or public-good interventions. The key move is the same: price the world if the action happens against the world if it does not.

Conditional Markets

When participants trade in these markets, they are not making unconditional bets. They are taking positions that only apply in a specific world.

In the current futarchy.fi experiment, this is implemented with conditional token markets:

  • Trades in the YES market only take effect if the proposal is approved.
  • Trades in the NO market only take effect if the proposal is not approved.

If the relevant world does not occur, the trade is reverted and funds are returned.

Tokens are useful here as an implementation technology. A conditional claim can be represented as a token even when the thing being evaluated is not a crypto-native token price. For example, a market could trade tokenized claims on a stock price, a revenue stream, an impact metric, a reliability score, or another objective.

What changes across designs is the objective being measured and the oracle that settles it. The token is the portable accounting object: the thing traders can hold, transfer, split, merge, and redeem. This is also how many prediction markets work in practice; even when the question is about an election or sports event, the traded positions are often tokenized claims.

What the Prices Represent

Because trades are conditional, prices can be interpreted branch by branch:

  • the price in the YES world reflects what traders think the objective is worth if the action is taken,
  • the price in the NO world reflects what traders think the objective is worth if the action is not taken.

The signal comes from comparing those prices. If the YES world is priced higher than the NO world, the market is saying the action is expected to improve the chosen objective. If the NO world is priced higher, the market is saying the action is expected to make things worse.

For example:

  • In a token-governance market, prices can estimate token value if a proposal is approved versus rejected.
  • In a company setting, markets could estimate share price or another business metric under different decisions.
  • In an agent team, markets could estimate whether a task will pass review, whether a change will reduce regressions, or whether a tool is reliable.
  • In a public-goods setting, markets could estimate marginal impact under each proposed intervention.

These prices reflect participants' beliefs, incentives, and information at the time the decision is made.

The future still matters: traders are pricing their expectations of future consequences into the market today. The important distinction is settlement. In the current futarchy.fi governance markets, settlement only needs to know which branch happened, not whether the proposal eventually turned out to be good. A different futarchy design could also include a later measured outcome, but that is an additional design choice, not required for the basic decision signal.

The Role of Measurement and Oracles

Every futarchy needs a credible measurement layer.

In the current futarchy.fi governance markets, an oracle is used only to determine which world occurred: whether the proposal was approved or not. The oracle does not decide whether the proposal was good. It only reports the decision outcome so conditional trades can be applied or reverted.

Other futarchy designs may need an oracle for the objective itself: revenue, usage, safety incidents, public-good impact, or another measured outcome. That measurement problem is often the hardest part of the design. If the metric is easy to manipulate, the market will optimize the wrong thing.

Why This Produces a Useful Signal

Futarchy markets let participants express beliefs with incentives attached.

If you believe an action improves the objective, you can take a position that pays in the world where that action happens. If you believe it makes the objective worse, you can take the other side.

In the current token-market implementation, this means:

  • If you believe approving a proposal will increase value, you can buy in the YES market.
  • If you believe approving it will decrease value, you can sell in the YES market.
  • If you believe rejection is better, you can trade the NO market instead.

In all cases, participants are taking positions that only apply if the corresponding world occurs.

As participants buy and sell in both markets, prices adjust to balance all of these views. The resulting YES and NO prices reflect the market’s collective assessment of which decision is expected to create more value, given the opportunity to express both positive and negative beliefs.

Participants are free to take positions in either or both markets; how positions combine is discussed in the trading section.

Where futarchy.fi Fits

Futarchy is still underexplored. Public implementations are early, and the most visible live experiments have been crypto-native because crypto provides tradable assets, transparent governance, and programmable settlement.

futarchy.fi should be read in that light: it is an experiment in conditional governance markets using token value as the current objective proxy. It is a working surface for learning about market creation, liquidity, conditional assets, oracle boundaries, and trader behavior. It is not the full scope of what futarchy can become.

Other crypto-native experiments, such as MetaDAO, show the same broader direction: markets can become a decision layer, not only a forecasting venue.

What Futarchy Is — and Is Not

Futarchy is:

  • A way to aggregate beliefs using markets.
  • A tool for evaluating decisions against a chosen objective.
  • A source of incentive-aligned signals for funding, governance, agent coordination, and execution.

Futarchy is not:

  • A traditional voting system.
  • A guarantee that decisions will succeed.
  • A substitute for choosing good objectives.
  • A reason to automate high-stakes decisions before the signal has earned trust.

How the signal is used — advisory, binding, or otherwise — is up to each project.

What’s Next