Game Theory and Strategic Interaction

Introduction

Managerial economics traditionally studies how firms, consumers and governments make choices under constraints. In many important decisions, however, the outcome does not depend only on an organisation’s own actions. It also depends on the actions anticipated from competitors, suppliers, regulators, employees, customers or negotiating partners. A firm deciding whether to reduce price, launch a product, enter a market or invest in capacity must therefore reason strategically.

Game theory is the systematic study of strategic interaction. It provides concepts and analytical tools for situations in which the payoff from one decision depends on the decisions of others. Its relevance extends beyond competitive rivalry. Managers use game-theoretic reasoning in bargaining, procurement, alliances, platform competition, innovation races, labour negotiations, cybersecurity and regulatory strategy.

The central managerial question is not simply, “What is my best action?” It is: “What is my best action, given what others may do, and what do they believe I will do?”

Learning Objectives

By the end of this chapter, the learner should be able to:

1. Explain the structure and assumptions of strategic games.
2. Distinguish dominant strategies, dominated strategies and Nash equilibrium.
3. Analyse simultaneous and sequential games using payoff matrices and backward induction.
4. Apply repeated-game reasoning to cooperation, retaliation and reputation.
5. Evaluate commitment, credible threats, signalling and entry deterrence.
6. Use game theory in rivalry, negotiation, collaboration and competitive strategy.
7. Recognise the limitations, ethical risks and institutional conditions affecting strategic models.

1. The Structure of a Game

A game normally contains four elements:

• Players: the decision-makers, such as firms, governments or negotiating parties.
• Strategies: the complete plans available to each player.
• Payoffs: the benefits or costs resulting from each combination of strategies.
• Information and timing: what each player knows and when decisions are made.

A strategy must be distinguished from an action. An action is a choice at a particular decision point. A strategy specifies what a player would do at every decision point that might be reached. This distinction is especially important in sequential games.

A game may be simultaneous when players choose without observing one another’s current choice, or sequential when one player moves before another. It may involve complete or incomplete information. In a game of complete information, the players know the relevant payoffs and available actions. In a game of incomplete information, they may be uncertain about a rival’s costs, quality, intentions or outside options.

The standard rationality assumption is that each player attempts to maximise expected payoff. This does not mean that people are perfectly calculating in practice. It means that the model represents decisions as purposeful responses to incentives. Behavioural factors, bounded rationality and organisational politics may produce departures from the model.

Payoffs and strategic interdependence

Suppose two competing firms choose either a high price or a low price. If both choose high prices, each may earn a substantial margin. If one chooses low while the other chooses high, the low-price firm may gain market share. If both choose low, price competition may reduce both firms’ profits. The payoff to a firm therefore depends on its own price and the rival’s price.

Payoffs can represent profit, market share, utility, political support, risk-adjusted value or a weighted combination of objectives. They need not be measured in monetary units, but the ordering of outcomes must be meaningful for the analysis.

2. Dominant Strategies and Nash Equilibrium

Dominant strategies

A strategy is strictly dominant if it gives a player a higher payoff than every alternative, regardless of what the other player does. A strategy is weakly dominant if it gives at least as high a payoff in every case and a strictly higher payoff in at least one case.

A dominant strategy is powerful because it does not require the player to predict the rival’s action. If both players possess dominant strategies, the combination is a natural prediction of the game.

Consider the following payoff matrix. The first number in each cell is Firm A’s payoff and the second is Firm B’s payoff.

For Firm A, low price yields 10 rather than 8 when B chooses high price, and 4 rather than 2 when B chooses low price. Low price is therefore strictly dominant. The same reasoning applies to Firm B. The predicted outcome is low price, low price, with payoffs 4, 4. Both firms would prefer the joint outcome high price, high price, but neither can obtain it unilaterally. This is a basic prisoner’s-dilemma structure.

Dominated strategies

A strategy is strictly dominated if another strategy gives a higher payoff in every possible circumstance. Rational players should not choose strictly dominated strategies, allowing analysts to eliminate them from a game. Iterated elimination can simplify complex games, although caution is required when dominance is weak or when players’ rationality is uncertain.

Nash equilibrium

A Nash equilibrium is a profile of strategies in which no player can improve their payoff by changing strategy unilaterally, given the strategies of the other players. It is a mutual best response.

For two players, strategy pair (s_A^*, s_B^*) is a Nash equilibrium if:

and

Here, u_A and u_B are the payoffs of players A and B, while s_A and s_B denote their strategies. The inequalities state that neither player can gain by deviating alone.

A Nash equilibrium need not be socially efficient, fair or desirable. Nor does it necessarily involve dominant strategies. It may be a pure-strategy equilibrium, in which players select definite actions, or a mixed-strategy equilibrium, in which players randomise according to probabilities.

Finding equilibrium in a payoff matrix

A practical method is to mark each player’s best response to every action of the other player. A cell is a Nash equilibrium if it contains a best response for both players. This method is appropriate for finite simultaneous games with observable payoffs.

Mixed strategies

Some games have no equilibrium in pure strategies. In such cases, players may randomise. Let player A choose action 1 with probability p, and player B choose action 1 with probability q. In a mixed-strategy equilibrium, each player chooses probabilities that make the other indifferent between the actions used with positive probability.

For example, in a matching-pennies-type game, predictable behaviour can be exploited. Randomisation prevents the rival from forecasting the action with certainty. In business, mixed strategies may appear in pricing promotions, inspection schedules, advertising timing, bidding and cybersecurity controls. Actual randomisation must be credible and sufficiently difficult for competitors to infer.

3. Sequential Games and Backward Induction

A sequential game represents decisions through a game tree. The first mover chooses an action, and the second mover observes that action before responding. The appropriate solution concept is often subgame-perfect Nash equilibrium, which requires strategies to form a Nash equilibrium in every subgame, including those that would not be reached along the equilibrium path.

The principal method is backward induction:

1. Analyse the final decision-maker’s best response at every possible decision node.
2. Substitute those responses into the earlier player’s problem.
3. Determine the earlier player’s optimal choice.
4. Report a complete strategy for every player and every relevant contingency.

Worked example: capacity entry

An incumbent firm chooses whether to accommodate or fight a potential entrant. If entry occurs, the incumbent can choose Fight or Accommodate. Assume the entrant first decides whether to Enter or Stay Out. Payoffs are as follows:

• Stay Out: entrant 0, incumbent 6.
• Enter and Accommodate: entrant 3, incumbent 4.
• Enter and Fight: entrant −1, incumbent 1.

After entry, the incumbent compares 4 from accommodation with 1 from fighting. Accommodation is optimal. Anticipating this, the entrant compares 3 from entering with 0 from staying out and enters. The subgame-perfect outcome is Entry followed by Accommodation, with payoffs −? No: the payoffs are entrant 3 and incumbent 4.

This example illustrates why an announced threat may not influence entry. If fighting is costly and accommodation is better for the incumbent after entry, the threat to fight is not credible. A threat is credible only when carrying it out is optimal when the decision point is actually reached.

Commitment

A commitment changes the future incentives available to a decision-maker. Examples include:

• investing in excess capacity before a rival enters;
• signing an exclusive distribution agreement;
• installing technology that makes reversal expensive;
• publicly announcing a price policy supported by contractual penalties;
• delegating authority to an agent with a narrower mandate.

Commitment can transform an otherwise non-credible threat into a credible one. However, commitment is costly and can reduce flexibility. A firm should compare the strategic benefit with the direct cost, opportunity cost and risk of escalation.

4. Repeated Games and Cooperation

A one-shot game may predict aggressive behaviour even when cooperation would benefit all players. In a repeated game, the same or similar strategic interaction occurs more than once. Future consequences can discipline present behaviour.

Let each period’s payoff be discounted by factor δ, where 0 ≤ δ < 1. A high δ indicates that future payoffs matter substantially; a low δ indicates strong short-term preference. The present value of a constant payoff π continuing indefinitely is:

The expression applies when the game continues indefinitely, the per-period payoff is constant and |δ| < 1.

Consider two firms that can cooperate for payoff C each period or defect for a one-period gain T, after which credible punishment gives each P per period. Cooperation can be sustained under a grim-trigger strategy when the value of cooperation is at least the value of cheating once and then receiving punishment:

Rearranging gives:

This condition assumes T > C > P, an indefinitely repeated relationship and a punishment strategy that is credible. Cooperation is more sustainable when firms value the future, cheating gains are small, punishment is effective and deviations are observable.

Repeated-game cooperation does not necessarily mean an illegal cartel. Cooperation may involve legitimate standard-setting, compatible technology, information sharing or supply-chain coordination. Explicit agreements to fix prices, divide markets or restrict output may violate competition law even if repeated-game logic predicts stability. Managers must distinguish economically sustainable coordination from unlawful collusion.

Reputation and forgiveness

A rigid punishment strategy may be too severe when errors, demand shocks or ambiguous actions occur. Tit-for-tat, forgiving strategies and graduated responses may perform better in noisy environments. Reputation is an intangible strategic asset: a firm known to retaliate may deter opportunism, while a firm known to honour agreements may attract partners. Yet reputational strategies can create escalation if signals are misinterpreted.

5. Signalling and Screening

When one party has information that another lacks, the interaction is a game of asymmetric information. A signal is an observable action intended to convey private information. A signal is credible when the cost or consequences of sending it differ across types in a way that makes imitation unattractive for the wrong type.

Examples include:

• warranties signalling product quality;
• education signalling ability in labour markets;
• large irreversible investments signalling commitment to a market;
• audited financial statements signalling transparency;
• introductory prices signalling confidence in future retention;
• certification signalling compliance or technical quality.

A signal is not automatically informative. If both high-quality and low-quality firms can copy it at the same cost, the signal may reveal little. Screening is the reverse process: the less-informed party designs choices that induce different types to reveal themselves. Insurance deductibles, subscription tiers and procurement qualification requirements are examples.

Pooling and separating outcomes

In a pooling equilibrium, different types choose the same signal, so the signal does not identify type. In a separating equilibrium, different types choose different signals, enabling observers to infer type. Equilibrium beliefs matter: observers must update beliefs consistently with the signals and the underlying incentives.

Managers should evaluate the net value of a signal:

The expression is a decision identity rather than a universal law. The benefit may include lower financing costs, increased demand or reduced monitoring. Signalling can be ethically problematic when it creates wasteful status competition or exploits consumers’ limited information.

6. Entry Deterrence and Competitive Strategy

Entry deterrence seeks to reduce the probability that a potential competitor enters a market. Common instruments include capacity expansion, patents, exclusive contracts, loyalty programmes, advertising, product proliferation and cost reduction. The strategic question is whether these actions alter the entrant’s expected payoff.

A potential entrant enters when its expected payoff is non-negative, subject to its risk preferences and financing constraints. In a simple model:

Entry deterrence is effective only if the entrant believes the incumbent’s response is both harmful and credible. Excess capacity may be a commitment if it is installed before entry and makes aggressive post-entry competition profitable. Merely claiming that the firm will fight may be cheap talk.

Limit pricing

An incumbent may charge a relatively low price to discourage entry. This strategy can work if the entrant interprets the low price as evidence of low incumbent costs and expects intense competition. It can fail if the entrant recognises the price as temporary or if the incumbent sacrifices more profit than the avoided entry is worth.

Predatory pricing and legal boundaries

Pricing below an appropriate measure of cost with the purpose of eliminating competitors may raise predatory-pricing concerns. Not every low price is predatory: price cuts may reflect efficiency, inventory clearance or legitimate competition. Managers must consider competition law, evidence of intent, recoupment prospects and consumer welfare. Game theory explains incentives; it does not authorise unlawful conduct.

7. Bargaining and Negotiation

Negotiation can be modelled as a bargaining game. Outcomes depend on surplus, outside options, patience, information and bargaining power. The reservation value is the least favourable outcome a party is willing to accept. The bargaining zone exists when the parties’ reservation values overlap.

Suppose a buyer values a contract at 100 and the seller’s minimum acceptable price is 60. The total negotiable surplus is 40, before considering transaction costs and risk. If the buyer has a strong outside option, its maximum acceptable price may be lower. If the seller has an alternative customer, its minimum acceptable price may be higher.

A credible deadline can improve bargaining power if the party is genuinely willing to walk away. An artificial deadline that is repeatedly extended damages credibility. Information disclosure can create value by reducing uncertainty, but revealing one’s reservation value may weaken one’s position. Negotiators should distinguish interests from positions: a supplier demanding a high price may actually require protection against volume risk or delayed payment.

Fairness, procedural legitimacy and long-term relationships matter. A purely distributive strategy may maximise a single transaction while destroying future cooperation. Integrative negotiation seeks trades across dimensions such as price, volume, timing, service, exclusivity and risk allocation.

8. Collaboration, Alliances and Platform Competition

Strategic interaction is not always adversarial. Firms may cooperate to develop standards, share research costs, build infrastructure or reach new markets. An alliance creates a repeated game involving trust, monitoring and governance.

Important design choices include:

• clearly defining each party’s contribution;
• allocating intellectual-property rights;
• specifying data access and confidentiality;
• establishing performance measures;
• defining dispute-resolution procedures;
• creating exit rights and termination consequences;
• preventing one partner from appropriating the other’s investment.

In platform markets, network effects make strategic interaction particularly complex. A platform’s value may increase as more users or complementary providers join. Early subsidies, exclusive arrangements and interoperability decisions can influence market tipping. These actions may produce efficiency benefits but can also exclude rivals and attract regulatory scrutiny.

9. Illustrative Case: The Northstar Logistics Market

Background

Northstar Logistics, an illustrative regional parcel company, serves business customers in a market where a new digital freight platform, Orbit, is considering entry. Northstar currently has a strong depot network and enjoys annual operating profit of 12 monetary units. Orbit must invest 5 units to enter.

Problem

Northstar must decide whether to expand depot capacity before Orbit enters. Expansion costs 2 units immediately. If Orbit enters and Northstar accommodates, Orbit earns 4 units after entry. If Northstar fights through temporary price reductions, Orbit earns −2 units and exits, while Northstar earns an additional 1 unit after accounting for the price war. However, fighting costs Northstar 3 units relative to accommodation. Orbit’s payoff from entry also depends on whether it expects accommodation or fighting.

Analysis

Without expansion, fighting is unattractive after entry because accommodation yields Northstar a higher post-entry payoff. Orbit therefore expects accommodation and enters if its net payoff is positive: 4 − 5 = −1, so it stays out. In this simplified case, the investment cost already makes entry unattractive.

Now suppose technological improvements reduce Orbit’s entry cost from 5 to 2. Orbit’s net payoff under expected accommodation becomes 2, so entry occurs. Northstar might consider expansion. If expansion costs 2 and makes fighting less costly—because additional capacity lowers Northstar’s marginal delivery cost—Northstar may be able to make fighting optimal after entry. The investment is then a commitment rather than mere excess capacity.

Northstar should compare the investment’s cost with the expected avoided competitive loss, while considering the possibility that Orbit will respond with a different business model, seek regulatory intervention or target an underserved segment. It should also test whether the capacity is genuinely irreversible and whether fighting would remain profitable under demand uncertainty.

Approach and solution

A robust strategy would combine:

1. scenario analysis for different entry costs and demand levels;
2. capacity investment only where it improves ordinary operating efficiency;
3. contractual and technological measures that reduce switching costs lawfully and transparently;
4. partnership or interoperability options with Orbit;
5. a contingency plan for accommodation rather than relying solely on deterrence.

Outcome

The case demonstrates that entry deterrence is not simply a declaration of aggressive intent. It depends on incentives after entry, the credibility of commitments, the entrant’s beliefs and the legal environment. An investment that is profitable only if a rival is excluded may be risky and legally sensitive; an investment that improves service efficiency has stronger strategic and economic justification.

Learning points

• A threat must be optimal when the threatened event occurs.
• Sunk investment can alter future payoffs, but it can also reduce flexibility.
• Entry decisions depend on expected payoffs, not only on absolute market size.
• Accommodation, partnership and deterrence should be evaluated as alternative strategic paths.

10. Limitations and Governance

Game theory simplifies reality. Payoffs may be uncertain, players may have multiple objectives, information may be noisy and managers may make systematic errors. Organisational incentives can also distort decisions: a divisional manager may prefer market share even when corporate profit falls.

Several limitations deserve attention:

• Bounded rationality: players may not calculate complex equilibria.
• Ambiguous payoffs: reputation, fairness and learning are difficult to quantify.
• Model uncertainty: the assumed game may omit an important competitor or constraint.
• Dynamic environments: technology, regulation and consumer preferences may change.
• Behavioural responses: retaliation, loss aversion and social preferences may affect choices.
• Multiple equilibria: the model may not predict which equilibrium will occur.

Managers should therefore use game theory as a disciplined framework for scenarios, not as a machine producing certainty. Sensitivity analysis, pilot tests, competitor intelligence and post-decision learning are essential.

Ethically, strategic reasoning must not be used to justify deception, collusion, exclusionary conduct, exploitation of vulnerable parties or manipulation of essential information. Artificial intelligence can support competitor-scenario analysis, negotiation preparation and detection of strategic patterns, but automated pricing or bidding systems may create coordination risks. Firms need human oversight, audit trails, data governance, competition-law review and controls against using sensitive competitor information.

Best-Practice Checklist

Before making a strategic decision, managers should:

1. Identify all relevant players, including regulators and affected stakeholders.
2. Specify actions, timing, information and payoffs.
3. Separate genuine strategies from unsupported threats.
4. Test for dominant and dominated strategies.
5. Identify Nash equilibria and, in sequential settings, subgame-perfect outcomes.
6. Examine whether commitments are costly, reversible or legally constrained.
7. Consider repeated interaction, reputation and learning.
8. Analyse alternative beliefs and uncertain payoffs.
9. Test cooperative and competitive options.
10. Review ethical, regulatory and stakeholder consequences.

Chapter Summary

Game theory analyses decisions in which outcomes depend on the actions and expectations of multiple parties. Its basic components are players, strategies, payoffs, information and timing. Dominant strategies provide an optimal choice regardless of the rival’s action, while Nash equilibrium identifies mutual best responses. A Nash equilibrium may be inefficient and need not involve dominant strategies.

Sequential games require attention to credible behaviour. Backward induction solves decisions from the final stage backward, and subgame-perfect equilibrium eliminates non-credible threats. Commitment can alter future incentives, but it may be costly and reduce flexibility. Repeated games explain how cooperation can be sustained through patience, monitoring, reputation and credible punishment.

Signalling and screening address asymmetric information. Signals are informative only when incentives make imitation difficult or costly. Entry deterrence depends on the entrant’s expected payoff and the credibility of the incumbent’s response. Game-theoretic reasoning also improves bargaining, alliance design, platform strategy and competitive analysis.

The framework is valuable but not mechanically predictive. Managers must account for uncertainty, bounded rationality, behavioural motives, changing institutions and legal constraints. The best strategic analysis combines formal models with evidence, scenario testing, ethical judgment and governance.

Key Terms

Game theory: The study of strategic interaction among decision-makers.

Player: A participant whose choices affect the outcome.

Strategy: A complete plan specifying actions under all relevant contingencies.

Payoff: The value a player receives from an outcome.

Dominant strategy: A strategy that performs at least as well as every alternative for every action of the rival.

Dominated strategy: A strategy that is always inferior to another available strategy.

Best response: The payoff-maximising action given another player’s choice.

Nash equilibrium: A combination of strategies from which no player benefits by deviating alone.

Mixed strategy: Randomisation across available actions according to specified probabilities.

Sequential game: A game in which decisions occur in an observable order.

Backward induction: Solving a sequential game by analysing later decisions first.

Subgame-perfect equilibrium: A Nash equilibrium that remains optimal in every subgame.

Credible threat: A threatened action that is optimal if the relevant situation occurs.

Commitment: An action that changes future incentives or restricts future choices.

Repeated game: A game played more than once, allowing future consequences to influence current behaviour.

Signalling: An informed party’s action intended to communicate private information.

Screening: A less-informed party’s attempt to induce information revelation.

Entry deterrence: Actions intended to reduce the likelihood or attractiveness of market entry.

Reservation value: The least favourable outcome a party is willing to accept in negotiation.

Bargaining zone: The range of agreements acceptable to all negotiating parties.

Review Questions

Short Answer Questions

1. What are the four fundamental elements of a game?
2. How does a dominant strategy differ from a Nash equilibrium?
3. Why can a threat be non-credible in a sequential game?
4. What is the role of the discount factor in a repeated game?
5. Distinguish signalling from screening.

Long Answer Questions

1. Explain how a payoff matrix can be used to identify dominant strategies and Nash equilibria.
2. Describe backward induction and illustrate why subgame perfection matters for managerial decisions.
3. Explain the conditions under which cooperation can be sustained in a repeated game.
4. Discuss commitment as a strategic instrument, including its benefits and costs.
5. Evaluate the strategic and regulatory risks of entry deterrence.

Analytical / Application Questions

1. Two firms repeatedly compete in a market. Cooperation produces 10 units per period for each firm, cheating produces 14 for the cheater and 2 for the other firm in that period, and punishment produces 4 per period thereafter. Derive the minimum discount factor needed to sustain cooperation under grim-trigger punishment.
2. Design a sequential game for a firm deciding whether to enter a market and an incumbent deciding whether to accommodate or fight. Specify payoffs that produce a non-credible threat and explain the subgame-perfect outcome.
3. A technology company is considering an expensive certification programme to signal product quality. Analyse whether the programme is likely to be separating or pooling, considering the different costs faced by high- and low-quality producers.

Multiple Choice Questions with Answers

Question 1

Which statement best defines a Nash equilibrium?

A. A strategy that is strictly better than every alternative in all circumstances

B. A strategy profile in which no player benefits from unilateral deviation

C. An outcome that maximises total industry profit

D. A strategy selected only when players move sequentially

Correct Answer: B. A strategy profile in which no player benefits from unilateral deviation

Question 2

In a sequential game, backward induction begins by analysing:

A. The first player’s preferences

B. The market’s historical trend

C. The final decision-maker’s best response

D. The socially efficient outcome

Correct Answer: C. The final decision-maker’s best response

Question 3

A threat is credible when:

A. It is publicly announced

B. It is costly to communicate

C. Carrying it out is optimal when the relevant decision point is reached

D. It reduces the rival’s current payoff

Correct Answer: C. Carrying it out is optimal when the relevant decision point is reached

Question 4

If a player has a strictly dominant strategy, that strategy:

A. Is optimal only when the rival cooperates

B. Produces the highest joint payoff

C. Is optimal regardless of the rival’s action

D. Must be a mixed strategy

Correct Answer: C. Is optimal regardless of the rival’s action

Question 5

In a repeated game, a higher discount factor generally means that a player:

A. Values future payoffs more heavily

B. Ignores future punishment

C. Has lower current costs

D. Cannot observe deviations

Correct Answer: A. Values future payoffs more heavily

Question 6

A separating equilibrium occurs when:

A. All types choose the same signal

B. Different types choose different signals

C. No player receives a payoff

D. Players randomise equally between all strategies

Correct Answer: B. Different types choose different signals

Question 7

Which action is most likely to serve as a commitment?

A. An unsupported verbal promise

B. A reversible private intention

C. A costly investment that changes future incentives

D. A prediction made by an analyst

Correct Answer: C. A costly investment that changes future incentives

Question 8

If both firms in a price game choose low prices because each benefits from undercutting the other, the outcome may be:

A. A Nash equilibrium that is jointly less profitable than mutual high prices

B. Impossible because low prices cannot be rational

C. A dominant outcome only if regulation prohibits pricing

D. Automatically socially efficient

Correct Answer: A. A Nash equilibrium that is jointly less profitable than mutual high prices

Question 9

Screening is best described as:

A. An informed party sending a signal

B. A less-informed party designing choices to reveal information

C. A regulator fixing market prices

D. A firm eliminating all uncertainty

Correct Answer: B. A less-informed party designing choices to reveal information

Question 10

Which factor generally makes cooperation easier to sustain in a repeated game?

A. A very large short-term gain from cheating

B. Poor observability of deviations

C. A strong valuation of future interactions

D. An inability to impose any response

Correct Answer: C. A strong valuation of future interactions