The Next Move Changes This One

You make an offer. The other person thinks about what you will do if they refuse it. You are already thinking about what they think you will do.

Welcome to game theory.

The name makes it sound as though someone has confused public life with entertainment. In this setting, a game is simply a situation in which the result of your choice depends on other people’s choices too. You cannot plan well by looking only at yourself.

That is why advice such as be confident can be so incomplete. Confidence may change how you present an offer. It does not tell you whether the other side can accept it, whether it believes your limit, or what happens after a refusal.

Nor does game theory provide a machine for predicting leaders. A model begins by simplifying. It chooses the players, the available moves, the information they possess and what they prefer. Change those assumptions and the result may change. The model is useful when it exposes a question we otherwise would have missed.

Let us build one slowly enough to see its moving parts.

Four possible corners

Imagine two fictional countries discussing a lawful shared-service project. Both available packages meet the required legal and rights standards. One package is basic. The other provides extra resources that make participation more attractive to the smaller country.

The smaller country can signal that it will participate on either package, or that it will participate only on the enhanced package. The larger country can offer the basic package or the enhanced one.

We give each country an invented preference score for each outcome. Higher means that country prefers the outcome. The scores do not measure money, happiness or moral worth, and a score for one country cannot be compared with a score for the other.

If the smaller country will participate either way and the larger offers the basic package, the scores are 1 for the smaller and 4 for the larger. If the larger instead offers the enhanced package, they are 2 and 3.

If the smaller country requires the enhanced package and the larger supplies it, the scores are 4 and 2. If the smaller requires it and the larger offers only the basic package, there is no agreement. The scores are 0 and 1.

The scores include preferences about discretion and precedent, as well as the package itself. In this invented case, the smaller country values having secured a recognized condition; the larger prefers retaining discretion over whether to provide extra resources. That is why the same material package can receive different scores after different negotiating paths. These assumed preferences are open to challenge. Change them, and the result may change.

The point is not to believe these preferences. The point is to ask what follows if these are the preferences.

Scroll sideways to see all columns. Keyboard: focus the table and use the left and right arrow keys.

Choice pair	Scores smaller / larger
Flexible participation + basic offer	1 / 4
Flexible participation + enhanced offer	2 / 3
Require enhancement + basic offer	0 / 1
Require enhancement + enhanced offer	4 / 2

Figure 2. Invented rankings for a lawful fictional project. Higher is preferred within each country’s scale. The numbers are neither measured geopolitical preferences nor comparable welfare totals.

If the smaller country expects a basic offer, it prefers flexible participation to breakdown: 1 is better than 0. If it expects an enhanced offer, it prefers the firmer position: 4 is better than 2.

The larger country also responds to expectations. If participation is flexible, it prefers the basic offer. If the smaller country will genuinely require enhancement, it prefers an enhanced agreement to no agreement: 2 is better than 1.

Two pure-choice combinations can therefore hold together. Flexible participation and a basic offer form one. A firm requirement and an enhanced offer form another. In each, neither side improves its score by changing only its own move while the other holds still.

That is the idea of an equilibrium. It does not mean the result is fair, stable forever or best for everybody. It means a specific set of choices fits the model’s incentives.

Now let one side move first

The grid leaves out timing. Add it, and the story becomes more demanding.

Suppose the larger country first makes a basic offer. The smaller country then chooses whether to accept. At that moment, if the only remaining outcomes are a basic agreement worth 1 or breakdown worth 0, it accepts.

The larger country can reason backward from that final choice. Why offer more if the smaller country will accept less when the moment comes?

A warning issued before the meeting may therefore fail if the other party expects it to collapse afterward. The smaller country can say it requires enhancement. But if it retains the option to accept the basic offer and prefers doing so to breakdown, the statement alone does not change the final decision.

This is the distinction between a threat and a credible commitment.

Thomas Schelling helped show how a party’s ability to commit can change bargaining. A commitment matters because it changes what others reasonably expect the party to do later. It can sometimes strengthen a negotiator by limiting discretion, but it can also create danger when circumstances change.¹

Suppose the smaller country has a genuine, publicly authorized rule for this project: participation requires the enhanced resources. The larger country now faces a different choice. Under our invented rankings, it prefers an enhanced agreement to breakdown. The rule can change the offer.

But read the next sentence carefully. If the larger country instead preferred breakdown to the enhanced agreement, the same rule would produce no deal.

A commitment is not a spell. It works through the other party’s incentives.

Tie the right things down

The practical lesson is not that small states should loudly trap themselves before every negotiation.

Some commitments deserve to be firm because they protect rights or legal authority. Others concern policy preferences that may need adjustment. A negotiator should know which kind of boundary it is defending.

A clear requirement for lawful treatment cannot be traded away because a model assigns a higher score to doing so. It defines which moves are allowed. An insistence on one particular supplier, payment date or administrative format may be negotiable if another arrangement meets the same public purpose.

Confusing these categories causes trouble. A government may become inflexible over a detail that does not matter while remaining vague about a protection that does. It may announce a deadline for political effect, then sacrifice an important safeguard to avoid missing it.

The best commitments often attach to an outcome or a condition: adequate funded capacity before an obligation begins; independent review where required; a limit that cannot be altered without proper authority. The method can remain open to discussion.

This gives the other side room to help solve the problem. It also makes the limit easier to explain. We are not asking you to admire our firmness. We are explaining what must be true for the cooperation to work.

That kind of discipline can preserve relationships because it reduces surprise. Partners know what is negotiable, what is not, and who can decide.

The clock has a vote

A country with an urgent need can have a weak position even when its case is excellent.

Ariel Rubinstein’s bargaining model studies alternating offers when delay is costly. Under its assumptions, patience affects the division that can be sustained. The result belongs to a precise model, not a universal law that the party who waits longest always wins.²

The everyday insight is familiar. A person who needs wages by Friday cannot bargain over a job in the same way as a person with a year’s savings. A supplier with perishable stock may accept terms it would reject next week if the stock could survive. Time changes alternatives.

For a country, reserves can buy time. So can reliable emergency finance, several suppliers, advance preparation and a public willing to accept a temporary cost for a clear purpose. These resources can matter before they are used.

Yet waiting is not always wise. A delayed decision may harm people, raise costs or allow an opportunity to disappear. A government can mistake indecision for strategic patience. It can also use the language of urgency to prevent proper scrutiny.

We need to ask what delay does to each side, and to the people outside the negotiation.

If waiting one month gives us useful information at low cost, waiting may improve the bargain. If the same month leaves people without a necessary service, delay has a serious price. The question is whether that price changes the decision and whether a temporary measure can reduce it.

Patience is most useful when somebody has prepared the means to afford it.

We do not know their whole mind

So far, our fictional countries knew the preference scores. Real negotiators do not.

The larger country may care mainly about speed. It may care about a visible success. It may fear setting a precedent. Different officials may disagree. A negotiator may have authority to promise less than their title suggests.

The smaller country faces uncertainty about those motives and constraints. The larger country faces uncertainty too: about local capacity, political support, legal obstacles and the reliability of its counterpart.

This is where information becomes a form of leverage.

A clear cost schedule can reveal that a requested limit reflects a real operating constraint. A documented approval process can show that a minister cannot lawfully settle a point alone. A credible alternative offer can demonstrate that refusal is possible. These signals are stronger when they are difficult to fake and connected to actual conduct.

A vague declaration of resolve may carry less information because anyone can make it.

Still, showing every weakness is not required for honesty. A government can disclose the facts needed to make a responsible agreement without exposing private case details or every tactical preference. There is a difference between keeping a negotiating reservation confidential and hiding a liability from the people who must pay it.

A useful question is who needs to know what, and for which decision. The public may need the spending limit. A court may need a confidential individual record. A negotiating team may need an internal fallback. These information needs differ, but none should vanish behind a single word such as secrecy.

A test can teach both sides

When uncertainty is important, a staged arrangement may reveal whether cooperation works before exposure becomes large.

A small initial project can test communication, funding, service quality and dispute resolution. It can show whether promised support arrives on time. It can identify the work that a planning document missed.

But a pilot deserves the name only if someone can stop or revise it.

If the first stage creates obligations that cannot be reversed, calling it a test does not make it safe. If people are already affected, their rights and needs continue even when the test is judged unsuccessful. If expansion happens automatically unless someone objects, the burden of proof has quietly shifted.

A proper stage has a decision at its end. What evidence will be reviewed? Who decides whether the next stage begins? What happens if the evidence is mixed? Which responsibilities remain if expansion stops?

For a programme involving human beings, simulated files and administrative exercises should test basic readiness before real people are made to discover missing steps. Later learning can improve a lawful operating system; it should not excuse beginning without one.

There is strategic value here. A staged proposal may be easier for the other side to accept because it reduces its uncertainty too. A country can offer cooperation that grows with demonstrated capability instead of choosing between immediate full commitment and permanent refusal.

Tomorrow can improve today

Many relationships repeat. Governments meet again. Firms bid again. Officials remember whether a promise was kept.

The prospect of future cooperation can make short-term opportunism less attractive. If taking advantage today destroys a valuable relationship tomorrow, restraint may become worthwhile. Repeated-game theory studies the conditions under which such incentives can support cooperation; it does not promise that repetition makes people fair.³

Some unequal relationships repeat precisely because the weaker party has nowhere else to go. A partner can exploit that dependence for years. The future then reinforces the imbalance rather than correcting it.

So ask what is actually at risk for the stronger side. Will a broken promise be documented? Can cooperation be paused lawfully? Are other partners watching? Is future participation genuinely valuable? Can a dispute be reviewed by someone whose decision matters?

Reputation has force when there is an audience that can respond and a record that audience can trust.

This is one reason public reporting can improve a bargain. It makes performance visible beyond the two officials who negotiated it. It helps a successor government distinguish an obligation from a favor. It gives potential partners evidence of how each side behaves.

Visibility alone will not overcome every imbalance. But invisibility can make opportunism cheaper.

Learn from the model and leave it behind

We have now changed the game several times. We began with a grid. We added timing, commitment, delay, missing information and future meetings. Each addition changed what a sensible strategy might look like.

That is the lesson.

There is no single posture called strong that works everywhere. A firm boundary may improve one offer and end another negotiation. Waiting may produce better information or greater harm. A public statement may create support or make compromise harder. Cooperation may build a reputation or deepen a dependency.

A capable government asks which situation it is in.

It also remembers what the model left out. Governments contain competing interests. People make mistakes. Officials may protect their careers. Domestic opposition may be principled, opportunistic or both. A change in circumstances can invalidate a careful forecast. Human dignity cannot be represented adequately by the government’s preference score.

The purpose of the exercise is to become more alert, not more certain.

When someone says that a small country has only one possible move, ask what assumptions create that conclusion. When someone promises that a dramatic refusal will force a better offer, ask what makes the refusal credible and what the other side would prefer to do instead.

Then ask the most useful question of all: what could we change before the next negotiation so that a better move becomes possible?

That is where game theory becomes public work. The reserve is funded. The alternative supplier is qualified. The regional service is organized. The rule is clarified. The case record is made reliable. The next move changes because the country has changed what it can do.

Sources and notes