Loading field map
How do a handful of firms interact when no single one controls the market, yet the market is too concentrated for perfect competition to apply? This question has driven market competition theory since the early nineteenth century. The subfield studies oligopoly—markets dominated by a few sellers whose decisions are interdependent. Over nearly two centuries, eight major frameworks have reshaped how economists model strategic interaction among firms. Each emerged from a limitation in the preceding approach, whether that limitation was an unrealistic assumption, a missing dimension of competition, or a failure to connect theory with empirical evidence.
The first formal model of oligopoly appeared in 1838, when Antoine Augustin Cournot published Researches into the Mathematical Principles of the Theory of Wealth. Cournot imagined two firms selling identical water from separate springs, each choosing a quantity to produce while assuming the rival's output would stay fixed. The resulting equilibrium—now called the Cournot–Nash equilibrium—showed that each firm's profit-maximizing output depends on the other's choice, and that the market price settles between the monopoly price and the competitive price. Cournot's framework introduced the core idea of strategic interdependence, but it assumed that firms compete in quantities, not prices.
Forty-five years later, Joseph Bertrand offered a sharp alternative. In an 1883 review of Cournot's book, Bertrand argued that firms in reality set prices, not quantities. If two identical firms compete by undercutting each other's price, he showed, the only stable outcome is that both charge marginal cost and earn zero profit—a result far more competitive than Cournot's. This became known as the Bertrand paradox: with just two rivals, price competition can replicate the perfectly competitive outcome. The Cournot and Bertrand models thus defined two poles of oligopoly theory. Cournot's framework predicted moderate market power; Bertrand's predicted its complete disappearance. Neither model, however, allowed firms to differentiate their products, a limitation that the next generation of theorists would address.
In 1929, Harold Hotelling introduced spatial competition in his article "Stability in Competition." He imagined two firms located along a line of consumers, each selling an identical good but differentiated by the consumer's travel cost. Hotelling showed that firms have an incentive to locate close together to capture the middle of the market—the principle of minimum differentiation—but that this clustering can destabilize price competition. Spatial competition relaxed the homogeneous-goods assumption by adding a geographic dimension: products were differentiated by location, and consumers chose the seller with the lowest delivered price (mill price plus transport cost). This framework made product differentiation a strategic choice, not a given.
Four years later, in 1933, Edward Chamberlin's Theory of Monopolistic Competition and Joan Robinson's The Economics of Imperfect Competition independently broadened the analysis of differentiation. Chamberlin modeled a market with many firms, each producing a slightly differentiated product, so that each firm faced a downward-sloping demand curve but free entry drove profits to zero. Unlike spatial competition, which typically analyzed a small number of firms choosing locations, monopolistic competition assumed a large number of firms and focused on variety rather than location. Both frameworks, however, shared a crucial insight: product differentiation gives firms some market power even when many rivals exist. Together, spatial competition and monopolistic competition transformed oligopoly theory from a debate about homogeneous goods into a richer analysis of how firms compete along multiple dimensions.
By the late 1930s, a different kind of pressure was building. The deductive models of Cournot, Bertrand, Hotelling, and Chamberlin offered elegant predictions, but they were difficult to test with the data available at the time. In 1939, Edward Mason and his colleagues at Harvard launched the Structure-Conduct-Performance (SCP) paradigm, which reoriented industrial organization toward empirical cross-industry analysis. The SCP view held that market structure—the number of firms, entry barriers, product differentiation—determines firm conduct (pricing, advertising, collusion), which in turn determines market performance (profits, efficiency, innovation). The causal arrow ran from structure to conduct to performance, and the key empirical prediction was that more concentrated industries would earn higher profits.
Joe Bain's 1951 study "Relation of Profit Rate to Industry Concentration" became a landmark test of this prediction, finding a positive correlation between concentration and profitability. The SCP framework gave antitrust authorities a simple rule of thumb: high concentration was presumptively harmful. Yet the SCP view treated entry barriers as structural features of an industry—economies of scale, capital requirements, patents—that were largely exogenous to firms' strategies. This assumption would soon be challenged by theorists who showed that barriers could be created and manipulated by incumbents.
The 1970s brought a wave of game-theoretic thinking into industrial organization, and one of its first applications was to entry deterrence. In 1977, Avinash Dixit published "Entry, Capacity, Investment and Oligopolistic Pricing," which showed that an incumbent firm could invest in excess capacity as a credible commitment to fight a potential entrant. If the incumbent builds capacity that would be profitable only if entry is fought, the entrant anticipates a post-entry price war and stays out. This insight—that entry barriers are not just structural givens but can be strategically created—marked a decisive break from the SCP tradition.
Strategic entry deterrence and commitment models drew on the logic of game theory, particularly the idea of credible threats. An incumbent's threat to cut price after entry is not credible unless the incumbent has made an irreversible investment that makes fighting profitable. By making barriers endogenous, this framework bridged the SCP view (which saw barriers as structural) and the full game-theoretic approach (which would soon model every aspect of strategic interaction). It also had direct policy implications: if barriers are chosen by incumbents, antitrust authorities must look beyond industry concentration to examine the strategic investments that sustain market power.
By 1980, game theory had become the dominant language of oligopoly analysis. The game-theoretic oligopoly framework, consolidated in Jean Tirole's 1988 The Theory of Industrial Organization, absorbed and superseded the SCP causal story. Instead of assuming that structure determines conduct, game-theoretic models treated structure, conduct, and performance as jointly determined by firms' strategic choices under different information conditions. The framework introduced equilibrium concepts such as subgame perfection and Bayesian Nash equilibrium, allowing economists to analyze dynamic competition, signaling, and asymmetric information.
This framework did not reject Cournot or Bertrand; rather, it generalized them. A game-theoretic model could nest Cournot quantity competition, Bertrand price competition, or spatial competition as special cases, depending on the assumptions about the strategy space and information. The key shift was methodological: the SCP paradigm had relied on cross-industry regressions, while game-theoretic oligopoly relied on formal modeling of strategic interaction. This change made predictions more precise but also more sensitive to assumptions—a trade-off that remains central to the field today.
In 1982, William Baumol, John Panzar, and Robert Willig published Contestable Markets and the Theory of Industry Structure, proposing a radical benchmark. A perfectly contestable market is one in which entry and exit are costless, so that potential competition can discipline incumbents even if the market is highly concentrated. In a contestable market, incumbents must price at average cost and earn zero economic profit, because any positive profit would attract hit-and-run entry. The theory challenged both the SCP view (which equated concentration with market power) and the game-theoretic emphasis on entry deterrence (which assumed that incumbents could profitably block entry).
Contestable markets theory did not claim that real markets are perfectly contestable; it offered a benchmark analogous to perfect competition but applicable to oligopoly. Its policy implication was that antitrust authorities should focus on entry conditions rather than concentration ratios. If entry is easy, even a monopoly may behave competitively. Empirical tests, however, found that few industries satisfy the assumption of costless entry and exit, and the theory's influence waned as a literal description of markets. Yet it remains a living reference point: it reminds economists that potential competition matters, and it forces game-theoretic models to specify why entry barriers exist rather than assuming them.
Today, no single framework dominates market competition theory. Game-theoretic oligopoly provides the most flexible toolkit for analyzing strategic interaction, and it is the default approach in academic research. Contestable markets theory survives as a benchmark for thinking about entry conditions, especially in policy discussions about deregulation and network industries. The Cournot and Bertrand models remain workhorses for teaching and for applied work where the quantity-versus-price distinction is central. Spatial competition and monopolistic competition continue to inform models of product differentiation, particularly in international trade and urban economics.
The leading frameworks agree on several points: firms' strategic interdependence is central; entry conditions matter for market outcomes; and product differentiation softens price competition. They disagree, however, on how much structure to impose. Game-theoretic models can generate precise predictions but require strong assumptions about information, timing, and rationality. Contestable markets theory offers a simpler benchmark but fits few real markets. The SCP paradigm's empirical approach—correlating concentration with profits—has been largely abandoned, but its policy question—when does market power harm consumers?—remains urgent. The result is a productive pluralism: different frameworks are better suited to different questions, and the field's strength lies in knowing which tool to apply when.