The unnamed people, scenes and conversations in this book are imagined to explain choices. They are not accounts of interviews, fieldwork or events from the author’s life. Historical cases and research findings have sources. Numerical examples identified as hypothetical use invented assumptions.
The order is ready to send. It has a supplier’s name, a price, a delivery date and a place for a signature. The machine will let a small workshop make a part it now buys from someone else. A customer has asked whether the shop can take a larger job. There is room for the machine near the back wall, once the old shelving comes down.
Then another announcement appears.
A government may change the rules on imported parts. Perhaps the change will raise a cost. Perhaps an exemption will cover this part. Perhaps the announced date will move. The owner reads the same paragraph twice. It says more information will follow.
The machine itself is available. The staff know what it does. Nothing has changed about the electricity socket or the strength of the floor. Yet the order stays on the desk.
An apprentice had hoped to learn the new work. A supplier had held a place in its production schedule. At home, one of the workers has been discussing a move closer to the shop. Extra hours would make the new rent easier to manage. Nobody has told that worker that the hours are cancelled. Nobody can promise them either.
The rule has not arrived, but decisions have begun to change.
Start there. If we begin with an argument about whether a government is right or wrong, we may miss the interval in which people must act without a settled answer. The government has a calendar for its decision. The workshop has a calendar for its order. The family has a calendar for the lease. These calendars can overlap without agreeing.
That interval is the subject of this book.
A decision cannot always wait for the answer
Imagine that you are the owner. The sensible response might seem obvious: wait until the position is clear. Why risk the business on a guess?
But the machine’s seller will not hold the delivery slot forever. The customer may give the job to someone else. The price may change. Staff may leave because another employer can offer a definite start. Waiting does not preserve the whole scene exactly as it was. It preserves some choices while allowing others to disappear.
Now imagine that you are the worker. Waiting could mean keeping your current home rather than signing a new lease. That may protect you from a rent you cannot meet. It may also mean losing the only suitable place near work, or continuing a long journey each day. You are not choosing between risk and no risk. You are deciding which risks you can bear, for how long, with the resources you have.
The owner and worker may even agree about the likely policy. Their different choices do not prove that one understands the news better. The owner has a credit line. The worker has a deposit to pay. The owner can postpone a machine while using an old one. The worker cannot postpone every part of family life. A decision that makes sense in one account can create a problem in another.
This is why advice to remain flexible needs more detail. Flexible how? Who will hold the stock, keep the spare room or work the uncertain shift? Who gets paid for keeping the choice open? If we cannot answer, flexibility may be a polite word for asking someone else to wait at their own expense.
I want us to take that expense seriously. People should have room to respond to a changing world. They should also have a say when someone else’s response leaves their own life suspended.
Several things can change at once
Suppose the government announces a new charge that will definitely begin next month. The owner knows the amount and believes the date. Imported parts will cost more. She may order some now. Her response comes before the charge begins, but the reason is an expected increase in cost, rather than doubt about what the rule will be.
Change the example. Officials are considering several possible charges and have not chosen one. The owner does not know which price to use in her calculation. Now uncertainty about the rule enters the decision.
Change it again. The charge is settled, but nobody knows whether customers will accept a higher price for the finished product. That is uncertainty about demand. Or the rule is clear on paper, but the office handling applications has not explained how it will classify the part. The uncertainty concerns administration.
These problems can arrive together. Calling all of them uncertainty may be convenient in conversation. It becomes a problem when we try to understand what would help. A clear explanation cannot undo a higher charge. A lower charge cannot resolve an unclear eligibility test. More time helps only if the time can be used and the information is likely to improve.
The owner should therefore ask what remains unknown. Is it the policy’s content, the date, the chance of an exception, the action of another government, or the customer’s response? A person can be fairly sure about one of these and doubtful about another. A news story that reports one answer may leave the decision almost unchanged.
Nor does a confident forecast settle an obligation. Someone may confidently predict that a proposal will fail. The proposal could still pass. Someone may be sure that a rule will be overturned. Until the legal position changes, people may have duties under it. The status of a rule and an opinion about its future are different kinds of information.
We will keep those differences visible. They save us from making a dramatic announcement do the work of a legal notice, or making a legal notice promise a future it cannot control.
The smaller choice inside the large one
Look at the unsigned order again. The owner has framed the choice as buy or wait. That might be too narrow.
Could she rent a machine for a short period? Test a smaller job with existing tools? Agree to supply part of the order rather than all of it? Ask the customer to share the cost of reserving capacity? Each possibility needs checking. Rental may be expensive. The smaller job may teach very little. The customer may refuse. A path that looks reversible can contain a deposit that is lost, a worker who has moved, or a promise that cannot fairly be withdrawn.
Still, these questions change the work. Instead of trying to predict every announcement, the owner examines what she would commit, what she could recover, and what the next step would reveal.
That is a more modest power than knowing the future, but it is a power we can use.
There is a danger here too. If we make clever private decisions the whole answer, we leave the people writing the rules out of the picture. A household should not need an advanced strategy for every rent payment. A shop should not need to monitor rumours throughout the night to understand a public deadline. Better individual judgment can reduce some harms while leaving the source of those harms untouched.
The book therefore has two tasks. One is to improve decisions made under changing rules. The other is to ask what those who change the rules owe the people who must live through the change. The tasks belong together. We need people able to act, and institutions worthy of the effort those people make.
A settled rule can still be a bad rule
There is an easy way to write about this subject: every change creates doubt; doubt discourages action; therefore governments should stop changing things. That argument would protect any existing arrangement simply because it exists.
Consider a different imagined workshop. Its ventilation is unsafe. Workers breathe dust because the owner has postponed an upgrade. A new safety standard will require better equipment. The owner complains that the date makes planning difficult.
We should examine the complaint. Perhaps a part is unavailable. Perhaps the rule gives conflicting instructions. Perhaps a phased change could protect workers sooner than an unrealistic deadline followed by widespread failure. Those are real questions.
But the workers have been living with a cost too. Keeping the old rule may spare the owner a new decision while exposing them to another month of harm. If we count only the cost of changing, we quietly assign a value of zero to the cost of continuing.
The same reasoning applies when a rule excludes people, protects an unfair advantage or fails to address a serious danger. Predictability is valuable because people use it to build lives. It does not excuse every arrangement around which a life or a business has been built.
A fair transition may require notice, assistance, compensation in some circumstances, or a different sequence. It also needs a reason to arrive somewhere better. We should not confuse a managed change with a permanent invitation to delay it.
This is a matter of judgment as well as measurement. Whose loss is counted? Who had a reasonable basis for relying on the old rule? Who was already paying for it? What would an extra month achieve? We cannot answer those questions by repeating that markets dislike uncertainty. People dislike unsafe air, unpaid wages and promises that never become services too.
The information we wish we had
The owner would like one number: the chance that the rule will change. With that number she could calculate a likely return, choose an option and get back to work.
Sometimes a defensible probability is available. Sometimes several informed people disagree. Sometimes the range of possible rules has not even been defined. In those conditions, a precise-looking number can hide a very rough judgment.
We can still reason. We can ask whether the choice changes if the probability is modestly higher or lower. We can examine a bad outcome without pretending it is the most likely outcome. We can identify a cost we cannot survive. We can decide which new information would justify another look. None of this requires a theatrical claim that everything is unknowable.
What matters is keeping the assumptions where we can see them.
Later, we will use a small calculation with invented numbers. A workshop can commit, buy information and wait, or take a smaller step. Different assumptions favour different choices. The calculation will not tell a real firm what to do. It will expose the reasons a recommendation changes, which is something a confident slogan rarely does.
We will also look at research. A newspaper index can show how often uncertainty is discussed. A business survey can show what managers say they plan to do. Company records can show investment that actually took place. A model can compare a recorded history with a carefully defined alternative. These sources answer related questions. They do not become interchangeable because the word uncertainty appears in all of them.
Learning to read their differences is part of learning to act. It helps us resist both the false comfort of a single prediction and the false sophistication of refusing to make any judgment at all.
Someone needs a date
The worker does not ask the owner for a lecture about world trade. He asks whether the extra hours will start next month.
An honest answer may be that she does not yet know. That can be the beginning of a responsible conversation, rather than the end of one. She can say which order is unsettled, when she will decide, and what hours she can guarantee in the meantime. She can avoid asking him to hold every other option open for her benefit. If she needs him to reserve time, she can consider what that reservation is worth.
The government faces a larger version of that duty. It may need room to negotiate or respond to new evidence. Yet a public authority can still distinguish a proposal from a requirement, publish the basis for a change, explain what remains open and give people a way to challenge a mistake. It can acknowledge the costs of a transition without pretending that all changes should stop.
The owner finally picks up the order. Whether she signs it is not yet the important thing. What matters is that the decision is no longer a test of whether she can guess the next headline. She needs to know what she is committing, which facts could change her choice, how long she can afford to wait and what her decision asks of other people.
The paper on the desk is only one part of that decision. The apprentice, the supplier and the family looking for a home belong in it too.
Source notes
This chapter develops original imagined examples and the book’s argument. It makes no claim that these events occurred or that the workshop represents a measured business population. Later chapters provide the research evidence, historical policy records and the fully specified hypothetical calculation.