Comment: Game Theory for the study of International Politics and its limitations. (2005, 20 Marks)
Game theory is the formal study of rational choice when one actor’s outcome depends on another’s choice. It was founded by John von Neumann and Oskar Morgenstern in Theory of Games and Economic Behavior (1944). Thomas C. Schelling’s The Strategy of Conflict (1960) carried it into strategic studies. It fits international politics because states under anarchy are interdependent decision-makers. It clarifies the structure of an interaction but predicts poorly.
The apparatus
- Elements: players, strategies (complete contingent plans), rules, pay-offs and information.
- Zero-sum games have a fixed total, and one side’s gain is the other’s loss. Non-zero-sum or mixed-motive games, which Schelling saw as typical of international conflict, let both sides gain or both lose.
- Solution concepts: minimax, the dominant strategy, and John F. Nash Jr.’s equilibrium, from which no player gains by moving alone. An equilibrium can still leave everyone worse off.
What it illuminates
- Chicken and brinkmanship. No dominant strategy exists, so commitment and a show of recklessness pay. Schelling’s threat that leaves something to chance explains the Cuban Missile Crisis, where Soviet withdrawal was the rational swerve. India’s strikes on Pakistan in May 2025, declared “non-escalatory” and followed by a ceasefire understanding on 10 May, can be read as two nuclear neighbours managing chicken.
- Prisoner’s dilemma. Defection is each side’s dominant strategy, so rational choice yields a collectively irrational outcome. This underlies arms races and the security dilemma (Robert Jervis, 1978).
- Trade wars as an iterated dilemma. In April 2025 the US–China tariff spiral reached 145% and 125%. The Geneva truce of 12 May cut the rates to 30% and 10%; the Busan summit of 30 October 2025 set a one-year truce, extended in September 2026 to 10 January 2027. Retaliation gave way to reciprocal, time-limited restraint.
- Stag hunt. Here the problem is assurance, not temptation, so the remedy is information. India and Pakistan have exchanged lists of nuclear installations every 1 January since 1992.
- Iteration. Robert Axelrod’s tournaments in The Evolution of Cooperation (1984) showed that tit-for-tat sustains cooperation when the “shadow of the future” is long. Robert O. Keohane’s institutionalism adds that regimes lengthen the shadow and expose defection.
- Two-level games. Robert D. Putnam (1988) showed negotiators bargaining abroad and at home at once.
Limitations
- Rationality. Herbert A. Simon’s bounded rationality and Jervis’s misperception show that leaders satisfice, misread and act under stress. Amartya Sen’s “Rational Fools” (1977) adds that reducing choice to self-interest ignores commitment and sympathy.
- The unitary actor hides bureaucratic and domestic bargaining.
- Pay-offs cannot be measured. Numbers for survival or honour import false precision. The analyst rarely knows which game is being played, and mistaking deadlock for a dilemma wastes diplomacy.
- Fixed preferences. Alexander Wendt and other constructivists argue that interests are formed by identity and interaction, so they are not given before the game.
- Relative gains. Joseph M. Grieco (1988) argued that states worry about who gains more, which limits the optimism of iterated cooperation.
- Unfalsifiable. A failed prediction can be rescued by redefining pay-offs after the fact.
- Normative hazard. Anatol Rapoport, whose tit-for-tat won Axelrod’s tournament, warned in Strategy and Conscience (1964) that strategic gaming rationalises nuclear risk. J. Ann Tickner (Gender in International Relations, 1992) argues that its rational actor universalises a masculine, context-free model of behaviour.
Conclusion
Game theory is a powerful heuristic. It shows why reasonable states fall into arms races and which remedy suits which structure: enforcement for dilemmas, reassurance for stag hunts, iteration for cooperation. It is not a predictive science. It is most useful when combined with history and decision-making analysis, which tell the analyst which game is actually being played.