Futarchy Explained

Futarchy is a proposed system of governance where citizens vote to define success, then betting markets decide which policies are most likely to produce it: the electorate agrees on a measure of national welfare, and prediction markets, not politicians, pick the policies forecast to raise that number the most.

The idea comes from economist Robin Hanson, who coined the word and laid out the design. It is a serious thought experiment about combining democracy with the forecasting power of markets, and it is now being tested at small scale in online communities. Whether you find it brilliant or alarming, it is one of the clearest ways to see what prediction markets are actually for.

The core slogan

Hanson compresses the whole design into a single line. It draws a sharp boundary between two questions that ordinary politics tangles together: what do we want, and what will get us there.

Values are the things we care about, such as prosperity, health, or safety. They are matters of preference, so we settle them democratically, by voting. Beliefs are factual predictions about which policy will best serve those values. They are matters of fact, so Hanson argues we should settle them the way markets already settle other hard forecasts: by letting people bet, and trusting the price.

The full proposal is set out in Hanson’s paper “Shall We Vote on Values, But Bet on Beliefs?”, published in the Journal of Political Philosophy, which remains the canonical source for the design.

How the mechanism works

Futarchy runs in two stages. First the public settles the goal, then the markets settle the method.

1. Define welfare. Elected representatives agree on a formal, measurable index of national welfare, for example a blend of GDP, life expectancy, leisure, and inequality. This is the number the whole system tries to raise.
2. Propose a policy. Anyone can put forward a concrete proposal, such as a new tax rule or an infrastructure plan.
3. Open two decision markets. For each proposal, the system runs a pair of conditional markets that forecast the welfare index: one assuming the policy is adopted, one assuming it is rejected.
4. Compare the forecasts. If traders price the welfare measure higher under adoption than under rejection, the market is predicting the policy would help.
5. Adopt by rule. The policy is enacted automatically whenever the adoption market forecasts higher welfare than the rejection market. No further vote, no lobbying, no committee.

The bets that turn out to be on the losing side of what actually happens are settled against reality, so anyone who traded on a bad forecast loses money and anyone who saw the outcome clearly is paid. Markets that get overruled (the rejected side of an adopted policy) are simply unwound, so traders are only ever judged on the branch that came true.

A concrete illustration

Suppose a country adopts a welfare index that is mostly GDP per person, adjusted for life expectancy. A legislator proposes a carbon tax.

• Market A prices the welfare index for five years out, assuming the carbon tax passes.
• Market B prices the same index over the same horizon, assuming it does not pass.
• Traders who believe the tax would grow the economy and extend lives buy in Market A and sell in Market B, pushing A above B.
• If A settles clearly above B, the rule enacts the tax. If B stays higher, the proposal dies.
• Whichever branch actually occurs is measured, and the trades on that branch pay out against the real welfare number.

Notice what the vote never decided: whether a carbon tax is good. Voters only decided that they care about GDP and longevity. The market decided whether this particular policy advances those goals. That division of labor is the whole point.

If the underlying tool is new to you, it helps to first read prediction markets explained, since futarchy is essentially a government built out of the same conditional markets.

Where it is being tried

No national government runs on futarchy, and none is close to it. The live experiments are all in software communities, where the stakes are lower and the rules are code.

Crypto projects and DAOs (online organizations governed by token holders) have been the natural testbed. Because a DAO already votes on-chain and controls a shared treasury, it can wire a market directly into its decisions: open conditional markets on a token’s price or a treasury metric under a proposal, and let the higher forecast decide. Some projects have used this style of “decision market” governance to choose between funding proposals, and it is often discussed under the label “governance by markets.”

These trials are small, and their welfare measure is usually something crude but easy to price, like a project’s own token value. That makes them a useful laboratory precisely because they sidestep the hardest question a real government would face.

Firms have run a milder version of the same idea for years. See corporate prediction markets for how companies already let internal markets forecast launch dates and sales, which is futarchy without the automatic-enactment rule.

The main criticisms

Futarchy is elegant on paper, and most economists who study it also see serious obstacles. Three come up again and again.

Defining welfare is a value judgment in disguise

The whole design assumes we can write down a single welfare number that the markets then optimize. But choosing what goes into that index (how to weigh growth against inequality, liberty, or the environment) is itself one of the deepest political disagreements there is. Bundle it into a formula and you have not removed the politics, you have hidden it inside a weighting the public may never scrutinize. Optimize the wrong measure and the machine will faithfully deliver the wrong outcome.

Manipulation and incentives to distort

When a market decision moves real power and money, actors have a strong motive to move the price rather than to forecast honestly. A wealthy interest could bet to make its preferred policy look good, betting that the resulting decision is worth more than the money it loses on the trade. Defenders answer that manipulators create profit opportunities for informed traders who correct them, but that defense weakens exactly when markets are thin.

Thin markets and rare, complex decisions

Markets forecast well when they are deep, liquid, and asking a crisp question. Policy markets are the opposite: many proposals, long horizons, fuzzy welfare measures, and few traders per question. Thin markets are noisy and easy to push around, so the very conditions that make a prediction market accurate are hardest to guarantee for the decisions futarchy cares about most.

These failure modes are not unique to futarchy. They are the general limits of the tool, covered in why prediction markets are accurate, which explains when a market’s price is trustworthy and when it is not.

Why it is worth understanding

You do not have to want to be governed by markets to get something from futarchy. It is a clarifying lens: it forces a clean split between what we value and what we predict, and it shows just how much of everyday argument confuses the two. Even critics tend to agree that separating those questions is a useful discipline, whatever mechanism you use to answer them.

The best way to build intuition for the forecasting half is to make predictions yourself and get scored on them, so you can feel the difference between a value you hold and a belief you can be wrong about.

Clutch lets you predict real news, politics, and sports with in-app credits and keeps score over time, so you can watch your own forecasting sharpen. Get the app and try betting on your beliefs.