A Promise That Can Change

Suppose instead that a higher charge on imported parts would make producing them in the shop more attractive. The owner hears that the charge will last. On that basis, a new machine looks worth buying. Then she hears that the charge is a bargaining tool. An agreement could remove it. Before the week ends, someone promises an exception for firms like hers.

All three statements might be sincere. They still leave her with three different plans for the same money.

She could ask which speaker is most likely to win. She could also ask a more basic set of questions. What has each speaker promised? Do they have the power to deliver it? What must happen first? And what happens to her order if the promise changes?

These questions can sound distrustful. In fact, they give a serious promise a chance to become useful. A promise to consider relief differs from a promise to grant it. An official may control the first and need a vote for the second. A lender deciding whether to finance the machine needs to know the difference. So does the worker deciding whether to train for it.

We should ask public officials to make these differences clear. We should also stop asking them for one assurance they cannot responsibly give: that no future evidence, emergency, election, or discovery of injustice will ever warrant a change.

What deserves to last

Imagine a town where an old license rule keeps new repair shops from opening. Owners who already hold a license can expect little competition. They have borrowed money and signed leases in that setting. Removing the restriction could make those investments less profitable.

The fact that they relied on the rule deserves attention. It does not settle whether the rule should remain. People denied a chance to open a shop have also lost something. Their loss is harder to display because there is no lease for the business they could not start.

If we measure fairness only by harm to existing plans, we give the people with plans a permanent advantage. A person kept outside the market must first become an owner to be heard, yet the rule prevents that very step.

My position is that public authority must retain the power to correct an unjust rule. The case for change should include the burdens of getting from here to there. It should also count the burdens imposed by staying here. A quiet year for license holders may be another lost year for everyone excluded.

This is a moral judgment, not a prediction that every reform will work. A government can misjudge the problem or choose a poor remedy. The answer is to examine the grounds for change, permit challenge, and take responsibility for the costs. Calling every revision a betrayal would protect errors along with legitimate expectations.

There are promises worth making firmly. An institution can commit to treating similar cases by the same published standard. It can commit to explaining a decision and correcting a factual mistake. Those promises concern how it uses power. They leave room to change what the power is used to do.

Why a sincere promise can fail

Economists Finn Kydland and Edward Prescott showed a problem with choosing policy afresh at each point in time. People act today on what they expect officials to do later. Once those choices have been made, the policy that looks best to officials can change. Their 1977 theoretical analysis explains why discretion can weaken an earlier plan, even with an agreed social goal. It does not establish that every fixed rule is wise.¹

We can take the warning seriously without handing the future over to the past. Before accepting an assurance, ask what will support it when keeping it becomes difficult. This is a test of the promise, rather than a test of whether the speaker seems decent.

Consider an imagined public training program. It offers places starting in January, but its funding ends in March. The course lasts six months. An honest director might expect the budget to be renewed. A worker who quits a job needs more than that expectation.

The director should identify the unfunded months. The public body could seek full funding before enrollment, shorten the course without damaging its value, or offer a stated refund and income plan if classes stop. Each route has costs. Hiding the gap simply leaves the worker to discover those costs later.

This example is a proposal about honest program design. It makes no claim that a refund can repair every loss. A worker may have given up a rare job, changed a care arrangement, or moved. The point is to make the limits visible before asking people to commit.

A stronger speech would not close the funding gap. A decision about resources might.

A threat has more than one audience

Policy can be a bargaining tool. Suppose a government threatens a new charge unless another government changes its conduct. In this imagined case, the threat is meant to make agreement more attractive than refusal. Whether it works depends partly on what the other side believes will happen after refusal.

That is one audience. The shop is another.

While officials bargain, the shop may pay for extra storage, revise prices, or lose an order it cannot quote with confidence. Those effects can occur whether the threatened charge arrives or disappears. The threat can be credible enough to alter a business plan while still being uncertain enough to make that plan difficult.

This does not prove that the threat was a mistake. It does mean that judging it only by the final agreement would leave out part of its cost. Citizens should be able to ask what happened during the bargaining period, who carried that burden, and whether the result justified it.

There is a genuine limit here. Publishing every possible concession could change the negotiation. Some information may need protection. Yet secrecy about a negotiating position is different from confusion about a rule already in force. A government can protect the former while making the latter clear.

It can say which charge applies now. It can distinguish a proposed date from an adopted one. It can name the office responsible for a formal update. It can say that an announced agreement still requires further action before businesses may rely on a new rate.

I would require this basic public account even when the negotiation remains private. People need to know which obligations govern their conduct. They should not have to infer them from the tone of a press conference.

We should be cautious with another argument: that confusion itself shows strength. Confusion may unsettle an opponent. It may also unsettle the people the government is meant to serve. The second effect belongs in the judgment, even if the first is real.

The exception that moves the uncertainty

An exception can solve a serious problem. Suppose the shop needs a part that no nearby producer can supply within the required time. Temporary relief might preserve repairs while another source is found. Whether relief is justified depends on the policy’s purpose and the evidence, not on the convenience of the applicant alone.

Now imagine the application rules are unclear. The owner can file, but cannot learn when a decision will come. She does not know whether other firms face the same test. She hears that persistence matters, though no one can explain what persistence means beyond making more calls.

The possibility of relief gives her a new choice. It also gives her another unresolved date. Ordering now could waste money if relief arrives next week. Waiting could lose the customer if it never arrives. An exception meant to ease the first problem has created a second one.

My proposed test is practical. The applicant should know the grounds for relief, what evidence is needed, who decides, and the expected decision time. The notice should explain whether a pending application changes any present obligation. An expected decision time must be labeled honestly if it is a service goal rather than a guaranteed deadline.

Comparable outcomes should be explained in a form others can use, with private details protected. If a firm gets relief because a part has no workable substitute, another firm using the same part should not need a personal introduction to discover that reason.

This would still allow judgment. A written test cannot foresee every case, and a rigid form can exclude a person whose problem is real. Officials should be able to explain why an unusual case meets the public purpose. The explanation is what makes that discretion open to challenge.

Relief also needs an ending. An extension can be justified, but a succession of unexplained extensions can turn a temporary measure into a lasting advantage. Each renewal should answer what has changed and why the original exit plan has failed.

A process people can inspect

There are existing legal examples of a commitment to procedure. In the United States, section 553 of the Administrative Procedure Act provides, within its scope, for notice of proposed rules, an opportunity to submit written views, and a statement of the adopted rule’s basis and purpose. It also contains exclusions and exceptions, including specified good-cause provisions. This is a bounded description of the statute in the 2024 United States Code, not a guide to which procedure governs any particular policy.²

The value for our discussion is the shape of the commitment. People can disagree about a rule and still insist that the authority making it follow the applicable process. A government need not promise that every commenter will get the result they want.

A right to comment can also be thin in practice. In our imagined town, a meeting at two in the afternoon may be easy for an association with paid staff and difficult for a mechanic who loses pay by attending. An online document may be public yet hard to read on the phone that a worker uses between shifts.

I would judge access partly by whether the people carrying the burden could actually contribute. That can mean plain explanations, more than one way to respond, and time to obtain the missing facts. The appropriate arrangement will vary with the decision and its urgency.

An institution should then show what it learned. If evidence changed a date, say so. If a requested exemption would defeat the rule’s purpose, explain why it was refused. A pile of submitted comments is evidence that comments arrived. It is not, by itself, evidence that the decision answered them.

Procedure needs a way to repair failure as well. If the wrong office supplied a misleading notice, a correction should reach the same people as the notice. If an application was rejected on a factual error, someone must have the authority and time to reconsider it. A promise without a repair route asks the injured person to carry all the risk of official error.

What can we honestly promise

The shop owner would still prefer to know the final charge. A clear procedure cannot give her that answer before it exists. We should admit the remaining uncertainty rather than sell good administration as a cure for politics.

What she can ask for is a statement with enough detail to support a decision. Here is the present rule. Here is the change under consideration. Here is who has authority to decide. Here is what remains unresolved. Here is the next formal step and where its result will appear.

For a larger change, I would add the conditions that could reopen the decision. Some can be named in advance, such as a failed safety test or a stated shortage of essential supply. Others cannot. When an unforeseen event requires revision, officials should explain the event and how it changes their earlier judgment.

That explanation should reach backward as well as forward. Who has already paid to comply? What can still be changed cheaply? Which harm needs quick relief? A revised policy may be justified while the way people were led into it still deserves criticism or repair.

The promise, then, is specific enough to test and modest enough to keep. Public power will remain able to act. The people affected will be told what has changed, given reasons they can question, and offered a fair chance to respond where the circumstances allow it. When speed makes that chance smaller, the duty to explain and review becomes more pressing.

Our shop owner may still postpone the machine. She may choose a smaller model or place the full order. But she can stop treating every remark as a possible instruction and begin using a public record that distinguishes words from decisions. That is a limited achievement. It is also a real one, especially for someone who has a business to run and cannot spend each day interpreting those who govern her.

Source notes