Advisory Futarchy

How can a public authority use forecasts without selling votes or surrendering lawful responsibility?

Futarchy separates two questions that public debate often mixes:

1. What outcomes and rights should govern this decision? Citizens and lawful democratic institutions answer this.
2. What is likely to happen under each option? Forecasters estimate this.

The market is an evidence instrument. It is not a voter, a legislature, a court, or proof that a system is decentralised.

Choose The Mechanism

Mechanism	What it does	Public authority
Forecasting tournament	Scores independent probability estimates without trading.	Receives evidence and decides.
Prediction market	Aggregates conditional estimates through priced contracts.	Receives evidence and decides.
Advisory futarchy	Runs forecasts against citizen-approved outcomes and requires the authority to explain its decision.	Retains full legal and budget responsibility.
Conditional delegation	Commits in advance to follow a signal inside narrow, reversible bounds unless a stated override applies.	Delegates a bounded choice but retains responsibility and stop power.
Binding futarchy	Makes the market signal execute or legally determine the decision.	Transfers effective control to the mechanism.

Use advisory futarchy by default in the public sector. Escalate beyond it only after comparable pilots show lawful operation, inclusive access, robust resolution, calibrated forecasts, and bounded harm. A forecast tournament may be safer than a market when liquidity, gambling law, unequal wealth, or financial access would distort participation.

Crypto rails are optional settlement infrastructure. A pilot can use play money, reputation, subsidised liquidity, or a tournament. A blockchain does not create equal citizenship, lawful authority, reliable data, or meaningful appeal.

Preconditions

Do not design the market until the decision passes the Civic Decision Readiness gate. At minimum:

• citizens or their lawful representatives approved the multi-metric setpoint;
• protected rights and affected subgroups are named;
• status quo and at least two feasible alternatives can be compared;
• the conditional question and resolution source are public and auditable;
• the legal decision-maker, budget owner, operator, auditor, and appeal routes are named;
• participation works without buying civic voice;
• the intervention is staged, reversible, and covered by a kill signal.

Operating Steps

1. Set values democratically. Approve the outcome setpoint, weights or decision rule, protected constraints, beneficiaries, and review mandate.
2. Qualify proposals. Public servants test legality, competence, funding, feasibility, conflicts, and data compatibility. Keep the status quo.
3. Freeze the forecast contract. State each conditional question, resolution date, data source, revision treatment, cancellation rule, and subgroup measures.
4. Open inclusive forecasting. Publish the option dossier. Declare conflicts, liquidity, position limits, access support, privacy controls, and manipulation monitoring.
5. Audit the signal. Test participation concentration, anomalies, sensitivity, model disagreement, and whether new information invalidated the contract.
6. Decide lawfully. The elected authority considers forecasts with legal, distributional, fiscal, delivery, and rights evidence. It publishes reasons, including any departure from the signal.
7. Execute in stages. Public servants deliver within the authorised bounds. Pause when a right, data source, integrity control, or kill signal fails.
8. Resolve and learn. Publish outcomes, data revisions, calibration, resident effects, appeals, and the consequence for the next decision.

Use the copyable Open Civic OS pilot canvas to produce the stable contract.

Proof Of Done

The mechanism must leave:

• a democratic-setpoint and protected-rights receipt;
• a qualified option dossier;
• a frozen forecast contract and version history;
• an access, privacy, liquidity, conflict, and integrity report;
• the forecast distribution, uncertainty, and sensitivity analysis;
• the authority's signed decision and reasons;
• the implementation mandate, review date, and kill signal;
• the resolution dataset, forecaster calibration, subgroup outcomes, appeals, and next-cycle correction.

Proof is not “the market was confident” or “the authority followed the price.” Proof is a dated comparison between the frozen contract and observed outcomes, with harms, exclusions, and revisions visible.

The method is done when all eight outputs exist, the authority and appeal body can retrieve them, and a cold reader can explain who set the values, who forecast the consequences, and who made the lawful decision.

Failure Modes

Failure	Control
The metric becomes the value	Citizens approve multiple outcomes and protected constraints; auditors publish gaming and subgroup effects.
Wealth buys influence	Separate equal civic voice from forecasting rewards; cap positions or use tournaments and subsidised participation.
Thin liquidity creates false precision	Publish depth and concentration; use intervals and HOLD the decision when the signal is fragile.
Insiders trade on privileged information	Apply disclosure, restricted-person, audit-log, and enforcement rules.
The authority launders responsibility through the market	Require a signed lawful decision and reasons, including departures.
The operator controls resolution	Use independent, predetermined sources and an appealable resolution process.
Averages hide harm	Freeze subgroup measures and kill signals before trading.
Digital identity excludes residents	Provide minimal-data, accessible, assisted, and offline eligibility routes.
Emergency powers swallow the protocol	Stop the mechanism; use existing emergency law and retrospective public review.

Explicit Exclusions

Do not use futarchy to determine:

• constitutional or human rights;
• criminal guilt, sentence, policing targets, detention, or surveillance of named people or groups;
• emergency coercion where delay or experimentation creates material danger;
• citizenship, voting rights, or access to justice;
• decisions with catastrophic, irreversible, or uninsurable downside;
• questions whose outcome data an interested party can secretly control;
• choices where affected people lack a practical path to participate or appeal.

Forecasting may still inform background analysis in some excluded domains. It must not become the decision rule.

Public-Sector Default

The Open Civic Operating System owns the democratic and institutional method. Prediction Markets explains the forecasting instrument. This page connects them.

The first worked design compares Auckland and Copenhagen housing. Both retain elected authority. Auckland tests whether forecasts can expose fragmented central-local and infrastructure dependencies. Copenhagen tests whether strong digital institutions can improve a housing choice without hiding renter, newcomer, affordability, or access disparities. Neither design is evidence that futarchy already improves government.

Changes my mind: comparable advisory pilots show that requiring public reasons does not prevent authorities or market operators from laundering responsibility through the forecast signal.

Questions

Which one forecast would materially improve a lawful decision without becoming the decision-maker?

• What protected value would a single target metric erase?
• Which affected group could be absent even when participation looks broad?
• What evidence would require the authority to reject the market signal?

Context

• Open Civic Operating System — place the mechanism inside rights, identity, service, appeal, and learning layers.
• Country Scorecard — test governability without ranking democratic legitimacy.
• Prediction Markets — inspect the forecasting instrument and its incentive risks.
• Tight Five — connect principles, players, platform, process, and performance before delegating action.

Next question: what is the smallest reversible public decision for which a forecast would change the authority's reasoning without displacing its responsibility?