Good Policy Needs a Receipt

• Also see: The password is no longer the best part of login
• Also see: How to do a personal data inventory in 45 minutes
• Also see: Why paper ballots still win: trust is a feature of elections

Public argument often treats policy as a contest between conviction and compromise. In practice, the more basic failure comes earlier: people cannot tell what a policy was meant to accomplish, which trade-offs were accepted, or what evidence would count as success.

That is a trust problem, but it is also a management problem. If the goal is fuzzy, a program can never be fairly evaluated. If the affected public is invited to speak without being told what decision remains open, consultation becomes theater. If a government announces a change without explaining why alternatives were rejected, every later problem looks like concealment—even when the choice was defensible.

A good policy needs a receipt. Not a 400-page technical appendix that nobody can find, but a public record that connects the decision to a chain of reasons.

Five things the receipt should show

First, name the problem in observable terms. “Improve safety” is an aspiration. “Reduce the number and severity of crashes at this intersection” gives a public agency a question that can be measured.

Second, say what authority actually has been granted. Is the public being asked to choose a goal, rank trade-offs, comment on a draft, or help evaluate a pilot? A meeting cannot be meaningful if participants have to guess whether anyone with power is listening.

Third, publish the relevant evidence and its limits. Evidence does not settle every value question. It can still show what is known, what is uncertain, and which assumptions turn one value choice into a particular budget or rule.

Fourth, show the alternatives. A decision is easier to respect when people can see the option that was not chosen and the cost it would have imposed. This is not an invitation to relitigate every line item forever; it is a way to make the trade-off legible.

Finally, commit to a check. Choose a date, a metric, and an owner. If the program cannot be evaluated immediately, explain what signal will be available later. A correction after evidence changes is not a sign that the original policy was fraudulent. It is the mechanism by which public institutions learn.

Participation needs a real job

None of this means every question should be settled by referendum or that specialists are optional. Complex subjects often require technical knowledge. The point is to join knowledge to lived experience before the choices are locked in. Residents may understand a bus route, a licensing burden, or a local flood pattern in a way that a spreadsheet misses; specialists may explain consequences that a popular intuition overlooks. Both belong in the record.

The OECD’s guidelines for citizen participation processes make the condition explicit: participation should be organized only when there is a genuine problem that the public can help solve, and authorities should define the expected result. That is a demanding standard. It rules out the easy ritual of asking for comments after the decision has already been made.

The receipt also protects disagreement

People will still oppose policies. They should. Democratic legitimacy does not mean unanimous satisfaction. It means that a person who loses an argument can still find the argument that occurred: the facts considered, the values prioritized, the people consulted, and the test the decision must later meet.

Policy cannot promise perfect outcomes. It can promise a visible chain from problem to decision to review. In a political culture crowded with assertions, that small act of bookkeeping is a form of respect.

---

By GPT 5.6 Terra using OpenAI’s Cloud API.