# Honest Politics Decision Standard

## Constitutional status

This is the controlling methodology for developing, assessing, approving, reviewing and ending Honest Politics policies. It replaces all earlier decision standards, gateways and methodology summaries.

It applies the Founder Doctrine. It does not replace political judgement, democratic authority, law or specialist professional standards.

## Governing principle

A complete record cannot rescue a weak decision. The method exists to make hidden assumptions, avoidable harms, weak delivery and concealed trade-offs harder.



Before supporting a major policy, Honest Politics will try to establish:

- what problem actually exists;
- why government should act;
- what realistic choices are available;
- what the evidence does and does not show;
- who benefits and who bears costs;
- which rights are engaged;
- whether the policy can be delivered;
- what could go wrong;
- how success or failure will be identified;
- what would cause the policy to change or end.

The work should be proportionate to the decision.

## The promising-idea screen and net-benefit test

Honest Politics cannot investigate every opinion or submission in full. Before formal policy work begins, an idea should have a plausible surface case: it must be relevant, material, lawful enough to examine and capable of being better than the present position.

Passing this screen is not approval. It means the idea deserves serious work, regardless of whether it came from inside the movement, another party, another country, an expert or a member of the public.

The net-benefit test asks:

> After rights and lawful constraints are respected, is the proposal clearly preferable to realistic alternatives once benefits, harms, costs, distribution, delivery, uncertainty, long-term effects and the people bearing the burden are considered?

## The twelve-stage sequence

### 1. Define the problem

State:

- the condition causing concern;
- who is affected;
- scale and severity;
- trend;
- causes and contributing mechanisms;
- geographical and territorial scope;
- what is known and uncertain.

Do not define the problem as the absence of a preferred solution.

Separate:

- symptoms;
- root causes;
- market failure;
- government failure;
- institutional failure;
- behavioural or social mechanisms;
- distributional and rights concerns.

### 2. Establish government justification and constraints

Before designing action, ask:

- Why should government act?
- Which government or institution has authority?
- What rights and legal duties apply?
- What is outside the legitimate scope of intervention?
- What minimum outcomes or safeguards are non-negotiable?
- What public, fiscal and delivery constraints are real?

Government action requires more than the existence of a social problem.

### 3. Define outcomes and the counterfactual

State the outcomes that matter and the period over which they matter.

Describe the most credible path without new action. The counterfactual is not automatically today's position held still. It should include likely behavioural, economic, demographic, technological and policy change.

Set:

- primary outcome;
- important secondary outcomes;
- guardrails;
- distributional outcomes;
- delivery outcomes;
- review horizon.

### 4. Generate a real option set

Consider genuinely different ways of achieving the objective.

The option set should normally include:

- no new intervention or current policy;
- improvement to existing policy;
- information, service or behavioural design;
- market or competition measures;
- regulation or taxation;
- spending or public provision;
- institutional or ownership change;
- local, devolved or targeted action;
- phased, temporary or experimental action.

Not every family is relevant to every problem. The record should explain why material options were included or excluded.

Do not compare a favoured proposal only with an implausible extreme. Do not discount an option because it came from another political party, profession, organisation or country. Credit material contributions.

### 5. Build the evidence case

Break the policy into claims.

For each material claim, identify:

- the claim;
- evidence required;
- available evidence;
- relevance to the UK and affected population;
- causal strength;
- uncertainty;
- conflicting findings;
- gaps;
- what further evidence would materially improve the decision.

Evidence quality depends on its fit to the claim. There is no universal hierarchy that makes one method best for every question.

Lived experience can reveal mechanisms, burdens and implementation failures. It does not by itself establish prevalence or causal effect.

### 6. Appraise consequences

Map the material consequences before compressing them into summary measures.

Include:

- benefits;
- public financial costs;
- economic and social costs;
- compliance and administrative burden;
- opportunity cost;
- transition effects;
- indirect and system effects;
- environmental and intergenerational effects;
- non-monetised consequences;
- fiscal effects;
- duration and persistence.

Use cost-benefit analysis, cost-effectiveness analysis, break-even analysis or multi-criteria comparison according to the decision.

Monetisation should be used where credible. It must not erase rights, dignity, constitutional effects, deep uncertainty or irreversible harm.

### 7. Test distribution, rights and fairness

Identify:

- who gains;
- who loses;
- who carries risk;
- whether effects differ by income, age, disability, geography or other relevant characteristics;
- whether a minimum standard is threatened;
- whether concentrated serious losses can be prevented, reduced, phased or mitigated;
- whether rights restrictions are lawful, necessary and proportionate;
- whether a less restrictive workable option exists.

A positive national average does not settle the decision.

### 8. Test behaviour, institutions and misuse

Ask how people and organisations may respond in practice.

Test:

- incentives;
- avoidance;
- gaming;
- fraud;
- moral hazard;
- adverse selection;
- principal-agent problems;
- metric manipulation;
- regulatory capture;
- administrative growth;
- stigma and take-up;
- displacement;
- technological adaptation;
- political misuse;
- hostile-successor abuse.

The policy should be assessed under realistic behaviour, not assumed compliance.

### 9. Test delivery

A policy is not ready because the idea is attractive.

Identify:

- responsible institution;
- legal powers;
- governance;
- staffing and skills;
- data;
- technology;
- procurement;
- suppliers;
- local and devolved roles;
- frontline workload;
- user journey;
- transition sequence;
- dependencies;
- cost and timetable ranges;
- operational risks;
- accountable owner.

A policy that cannot be delivered credibly must be redesigned, staged, tested or rejected.

### 10. Test legitimacy and challenge

Establish what public opinion, affected groups, professionals and independent reviewers reveal.

Distinguish:

- representative opinion;
- open submissions;
- affected-group experience;
- expert evidence;
- organised campaigning;
- moral and constitutional objections.

Expose major proposals to serious challenge before the decision is fixed.

The strongest reasonable objection should appear in the decision record.

### 11. Decide, design learning and state conditions

The decision-maker should choose among:

- proceed;
- proceed with conditions;
- test and learn;
- redesign;
- pause;
- reject;
- retire an existing policy.

The decision statement should explain:

- preferred option;
- why it is better than realistic alternatives;
- principal benefits;
- main costs and losers;
- rights judgement;
- major uncertainty;
- delivery confidence;
- political judgement;
- conditions;
- what would change the decision.

Where uncertainty is decision-relevant, build learning into implementation.

### 12. Monitor, evaluate, adapt and end

Define before implementation:

- baseline;
- outcome measures;
- process and delivery measures;
- guardrails;
- data source;
- evaluation question;
- evaluation design;
- review date;
- prediction or forecast entries;
- thresholds for continuation, amendment, expansion or stop;
- policy owner;
- publication plan.

Monitoring shows what is happening.

Evaluation examines whether, how and why the policy caused outcomes and whether it represents value for money.

A policy should be amended or ended when evidence, circumstances, legality, delivery or priorities no longer justify it.

## Four scrutiny levels

### Level 1: routine

For low-cost, reversible, low-rights-impact decisions.

Minimum:

- short problem and objective;
- options considered;
- main evidence;
- cost and delivery check;
- owner and review.

### Level 2: standard

For meaningful policy changes with limited systemic or rights risk.

Minimum:

- complete concise decision record;
- evidence and option comparison;
- distribution and rights check;
- delivery plan;
- monitoring and review.

### Level 3: major

For high-cost, nationwide, coercive, difficult-to-reverse or institutionally significant policies.

Minimum:

- full technical record;
- independent analytical and subject challenge;
- public or affected-group engagement where relevant;
- model and forecast assurance;
- implementation readiness;
- evaluation plan;
- published decision statement.

### Level 4: exceptional

For constitutional change, severe rights impact, catastrophic downside, major national-security implications or highly irreversible decisions.

Minimum:

- enhanced legal and constitutional review;
- multiple options and scenarios;
- deep-uncertainty and hostile-successor testing;
- independent challenge;
- explicit democratic-authority route;
- staged implementation where possible;
- strong stop, sunset or review controls.

## Five core records

The methodology uses five core records rather than a large collection of overlapping gateways.

### 1. Policy brief

Problem, objective, authority, status and scope.

### 2. Evidence and options record

Claims, sources, alternatives, appraisal and uncertainty.

### 3. Rights, distribution and risk record

Rights, winners and losers, behavioural response, misuse and major risks.

### 4. Delivery and learning plan

Implementation, monitoring, evaluation, review and exit.

### 5. Decision statement

The authorised judgement and reasons.

Registers such as predictions, corrections, interests and versions sit across the system but do not duplicate these records.

## Emergency route

Urgency may shorten the process, but it does not eliminate responsibility.

Emergency decisions require:

- authority;
- immediate objective;
- principal evidence;
- rights and harm check;
- time limit;
- accountable owner;
- monitoring;
- scheduled reconsideration;
- clear route to continuation or expiry.

## Final principle

The method exists to improve judgement, not to imitate certainty.

A complete form cannot rescue a weak policy. A strong decision should still be understandable in plain English.