Oskar Morgenstern, “Die Anwendung der Spieltheorie in der Wirtschaftswissenschaft” (1962)

Morgenstern’s essay is a methodological defense of game theory as more than a mathematical curiosity or auxiliary calculating device. He begins by conceding its autonomous scientific standing, then asks whether economic situations themselves are often strategic games rather than ordinary maximization problems.

Die Spieltheorie ist von wissenschaftlichem Interesse als ein Zweig der Mathematik und, schon als solcher, unabhängig von irgendwelchen Anwendungen.

English translation: Game theory is of scientific interest as a branch of mathematics and, as such, is already independent of any applications.

The essay first reviews familiar points of contact: utility under risk, the relation between two-person zero-sum games and linear programming, and the usefulness of formal analysis in clarifying complex interdependence. But Morgenstern sharply limits optimization models. Linear planning works where variables are controlled by the planner or can be statistically absorbed; it fails when another purposive agent controls a decisive variable and reacts to one’s action.

Offensichtlich gibt es solche Fälle: ausschlaggebend für die Verbindung zwischen Spieltheorie und Wirtschaftswissenschaft ist die grundlegend wichtige Tatsache, daß hier die normalen Maximumannahmen (mit oder ohne Beschränkungen) nicht anwendbar sind.

English translation: Obviously such cases exist: decisive for the connection between game theory and economics is the fundamentally important fact that here the normal maximum assumptions (with or without constraints) are not applicable.

This boundary defines the essay’s main claim. The firm that plans internally may resemble a maximizer, but the firm that sets price, advertises, threatens entry, negotiates, or anticipates retaliation is in a game. Morgenstern’s examples—railroad share struggles, advertising wars such as Lestoil against Procter & Gamble, Coca-Cola against Pepsi-Cola, cartel negotiations, Austrian tubes, and European mirror glass—are not anecdotes added for color. They are evidence that economic action is structured by counteraction, secrecy, bluff, public justification, coalition, compensation, and pressure.

From this perspective, received price theory appears too thin. Walrasian equilibrium treats actors as if they face fixed data, while duopoly and oligopoly cannot be understood by merely extending the same maximum calculus. Aggregate demand and supply curves conceal the number, weight, timing, and possible coalitions of participants. Morgenstern therefore regards the familiar crossing of curves as a special construction, not a universal description of markets.

The essay is not anti-formal. Morgenstern admires exact mathematical work when its assumptions are explicit, as in von Neumann’s model of balanced expansion. His criticism falls instead on mechanical metaphors of “forces” and “equilibrium” when they displace the lived structure of business conflict. For him, rules, information, concealment, and possible coalitions are not impurities but central economic facts.

Es gibt kein Geschäft ohne Geschäftsgeheimnis.

English translation: There is no business without a business secret.

This leads to the essay’s strongest preference for cooperative n-person game theory. Non-cooperative equilibrium analysis is useful, but it remains close to the classical image of isolated competition. The more radical economic promise lies in treating cooperation, cartelization, mergers, wage bargains, compensating payments, and coalition values as formal objects. Cooperation begins inside the firm and extends through the whole economy; it is not an accidental exception to competition.

Morgenstern’s discussion of “Macht” follows from the same argument. Marginal productivity theory and ordinary maximization leave no place for power, exploitation, or threat, because they suppress the strategic dependence of one actor on another’s decision. Game theory can compare coalitions of unequal strength and can model board struggles, shareholder alliances, union bargaining, and distributive conflict as problems of imputation rather than automatic market reward.

Diese beiden Auffassungen sind unvereinbar; unter den Annahmen dieser Theorie kann „Macht“ nicht vorkommen. Wenn man aber ihre Existenz zugibt, dann muß die herrschende Theorie weichen.

English translation: These two views are incompatible; under the assumptions of this theory, "power" cannot occur. But if one admits its existence, then the prevailing theory must give way.

The closing pages defend the vocabulary of “solution” and “Zurechnung” against the objection that it is artificial. Morgenstern turns the charge around: economists have inherited eighteenth-century mechanics, while businesspeople already speak naturally of games, fights, rules, maneuvers, bluffs, and countermoves. The essay’s enduring importance is its insistence that economics must model interaction, not only optimization, and that cooperative game theory is indispensable wherever secrecy, compensation, coalition, and power shape economic outcomes.