Skip to content

Hill – Mankiw 9th Edn Chapter 17 – Oligopoly

A commentary on Mankiw 9th Edn Chapter 17 – Oligopoly (Mankiw 9th edition)

Mankiw, N. G. (2021) Principles of microeconomics (9th ed.)
Principles of economics (9th ed.)
Mason, OH: South-Western Cengage Learning.

Rod Hill

University of New Brunswick, Saint John campus

Saint John, New Brunswick, Canada


Chapter 17 – Oligopoly

Here are some things to consider when reading this chapter.


  1. Mankiw’s Duopoly Example

Section 17-1 describes a duopoly of two firms selling spring water produced at constant marginal cost, assumed to be zero in this case. There are no fixed costs. Mankiw doesn’t mention its origin, but the example is taken from a book published by the French economist Augustin Cournot in 1838 (Cournot, 1838 [1960]). It’s the original analysis of what is termed ‘Cournot competition’; each firm decides how much to produce to maximize its profits, based on an assumption about what the other firm is producing. The products produced are identical (or homogeneous), so demand for them can be added together to get market demand. The total quantity produced determines the market price.

An equilibrium occurs when neither firm has any incentive to change the quantity it produces given its correct expectation about what the other firm will produce. The result is often termed a Cournot-Nash equilibrium when it occurs in the context of Cournot’s model.

Incidentally, this parallels the story Mankiw tells at the beginning of the chapter about the market for tennis balls in the United States. Producers “determine the quantity of tennis balls produced and, given the market demand curve, the price at which tennis balls are sold” (p. 335). However, for a market demand curve to exist, tennis balls would have to be perfect substitutes for each other. But tennis balls are differentiated by brand names and balls of different qualities, which makes it unlikely that buyers see them as perfect substitutes. With no market demand curve, a different model seems necessary in that case.

(i) “The Equilibrium for an Oligopoly”

This is the title of Section 17-1c. The word ‘the’ implies that an oligopolistic industry has a unique equilibrium, namely that described by Cournot’s model. Yet this is not the case, even in this simple duopoly setting.

Another French economist, Joseph Bertrand, proposed an alternative to Cournot’s model in a famous paper in 1883. He showed that if the producers of spring water make strategic choices about prices, instead of about quantities, the resulting equilibrium is different from Cournot’s.

In Bertrand’s model, the two firms also have no fixed costs and constant marginal costs. They set prices simultaneously and supply whatever is demanded at that price. Consumers buy from the firm with the lowest price. If firms set the same prices, they share the market equally.

Because the firm at the lowest price gets the entire market, each firm has an incentive to charge a price lower than that of the other firm. As a result, in the Bertrand-Nash equilibrium both firms set price equal to marginal cost. (Because there are no fixed costs and marginal costs are constant, any price below that would result in losses.) Neither firm has an incentive to change its pricing strategy given what the other firm is doing.

The result is the competitive equilibrium, even with just two firms. This contrasts with the Cournot model described by Mankiw, in which sellers’ prices only approach marginal cost as the number of firms gets progressively larger (pp. 339-40).

By briefly setting out Bertrand’s model, whose logic is easily explained, Mankiw could have introduced the idea that there are other possible models of oligopoly, even in this simple static situation, and that judgement is required as to which is most appropriate in a particular setting. In many markets, firms may also make strategic choices about things other than price or the quantity produced. For example, as in the case of tennis balls, producers may make decisions about product characteristics and advertising, differentiating their product from those of their rivals. In other markets, firms may make decisions about location, warranties to offer, the quality of customer service, and so on.


  1. Rivalry or cooperation among oligopolistic firms?

The Cournot model, as well as the Bertrand model described above, assume that firms don’t cooperate with each other. Mankiw’s text is typical of introductory texts; it explains the prisoners’ dilemma, stressing the difficulty of cooperation if firms choose their strategies simultaneously and must stick with them. If firms interact repeatedly, as is typical in most markets, the prisoners’ dilemma illustrates the temptation to deviate from a cooperative arrangement. This theme is reiterated in the chapter’s conclusion.

Then one section (17-2e, pp. 346-7) describes how “in a game of repeated prisoners’ dilemma, the two players may well be able to reach the cooperative outcome”, for example through the use of the tit-for-tat strategy by which non-cooperators are punished.

What are students to conclude about the frequency with which oligopolistic firms are able to evolve cooperative behaviour? The bulk of the discussion is about the difficulties of cooperation, yet this final section, describing a more realistic market setting, leaves open the possibility that cooperation/collusion could frequently evolve.

Mankiw describes only one kind of collusion: that in which firms make explicit agreements to cooperate, effectively forming cartels. This is overt collusion, which is illegal and can be prosecuted with sufficient evidence.

As Levenstein and Suslow describe, firms can negotiate about prices, market shares, “terms of sale, advertising, transport costs, and production capacities. … Firm asymmetries and changes in firm’s costs can make these negotiations challenging” (2008, p. 2). They report evidence, based on prosecuted cases, that the “average duration of cartels measured over a range of countries and time periods is between five and seven years”. Many last just a year or two, but the average is pulled up by those which last 10 or more years.

Surprisingly, Mankiw fails to mention a second kind of collusion: tacit collusion, in which firms evolve an understanding about how to behave cooperatively without communicating directly with each other. In contrast with overt collusion, tacit collusion is legal, so antitrust law does not apply. (In the United States, the Federal Trade Commission (FTC) can intervene if a firm makes a public announcement about future changes in prices and quantities that can be interpreted as an invitation to other firms to collude (Ballou 2021, p. 242).)

For example, tacit collusion can be maintained if one firm is understood by others to be the price leader, helping to coordinate price changes throughout the industry while keeping them above competitive levels. Other firms could then follow its lead as they change prices in response to changing industry demand or cost conditions. Antitrust authorities cannot deal with such ongoing tacit collusion because there is no explicit communication between the firms.

Unlike with overt collusion, no estimates exist about how long the firms might be able to tacitly collude. Nor can anything be said about how frequently tacit collusion occurs. After all, even if firms are behaving competitively, they will likely be changing their prices in the same way in response to changes in demand and costs. However, careful empirical work can uncover industry behaviour that is consistent with a tacitly collusive oligopoly, as in the gasoline market study cited in part 3(i) the Commentary on Chapter 14.

Some worry that tacit collusion could be facilitated now that prices in some markets are determined not by people but by computer algorithms. This is happening as sellers move online or as sellers adopt ‘dynamic pricing’. With dynamic pricing, algorithms change prices frequently according to the most recent information about market conditions. This can include information about the prices currently being offered by other firms, whose prices are also being set by algorithms. The algorithms could be explicitly written to facilitate tacit collusion or may be capable of learning how to do so (Ezrachi and Stucke, 2020).

Such concerns prompted Brendan Ballou (2021) to propose that the FTC implement a ‘no collusion’ rule. Under such a rule, a firm couldn’t raise its price simply because another firm in the market has raised its price. This would deal with price setting algorithms that consider only on the prices of other firms in the market. It would also address the common practices he cites. “For instance, when American Airlines announced that it would charge passengers for the first checked bag, United and US Airways quickly followed suit. … Conversely, when one carrier instituted a new fee for in-flight soft drinks and competitors failed to follow suit, the carrier retreated” (Ballou 2021, p. 223).


  1. Antitrust Policy

American antitrust policy has been influenced by the Chicago School, a group of economists and law professors based at the University of Chicago. According to Herbert Hovenkamp and Fiona Scott Morton, they “were libertarians who were committed on ideological grounds to less intervention by the state” (2020, p. 1847). This was accompanied by a faith in the self-correcting nature of markets. Cartels were seen as inherently unstable. Barriers to entry were downplayed, so that even monopolies would have to contend with competition in the long-term. Mergers were beneficial because they reduced costs and ultimately prices to consumers (as explained by Mankiw in Chapter 15, p. 309).

They write: “The Chicagoans embraced economics when it would achieve their anti-enforcement ends, but largely ignored its advances in theory and empirical technique after 1970 because those tools sometimes proved that anticompetitive conduct had occurred, and that enforcement was needed” (2020, p. 1853).

University of Chicago economist George Stigler portrayed regulation as the result of “political struggles under which different interest groups competed to see who could benefit the most from a particular government policy” (2020, p. 1854). Hence his claim that regulators are ‘captured’ by the regulated industry, so that regulations benefit the industry, not the public.

Hovenkamp and Morton describe how the Chicago School’s noninterventionist message is amplified by firms that would profit by it. They fund conservative institutions who organize “education and influence programs targeting academics and the judiciary” (2020, p. 1851). They contend that the result is a gap between relatively weak antitrust enforcement in the United States and current economic theory, which provides a rationale for more enforcement in the face of sophisticated strategies by oligopolistic firms. In an ironic inversion of Stigler’s story, they write the Chicago School has become “an economically outdated but nevertheless powerful tool of regulatory capture” by those “who stand to profit from nonintervention” (2020, p. 1844).

This background about the Chicago School is useful in considering Mankiw’s discussion of public policy towards oligopolies. In examining “controversies over antitrust policy”, he repeatedly emphasizes the position of the Chicago School.

(i) Resale/Retail Price Maintenance (RPM)

The setting of sellers of a minimum price for producer’s product (an example of RPM) is currently illegal under American antitrust law. Mankiw sets out two arguments by “some economists” who defend RPM (p. 349).

The first argument is that a producer wanting “to exert its market power … would do so by raising the wholesale price [that it charges to retailers] rather than controlling the resale price.” While this sounds persuasive, as a general statement it is not true. It overlooks circumstances in which (to give one example) a producer can increase both its own profits and those of retailers by setting both wholesale and retail prices. This prevents individual retailers from using their bargaining power to negotiate lower wholesale prices which, in turn, would result in retail prices lower than those which would maximize the joint profits of producers and retailers (O’Brien and Shaffer 1992).

Mankiw offers an argument in defence of RPM. Suppose the manufacturer believes that point-of-sale service provided by the retailer will increase demand for the product (e.g. providing product information or demonstrations, longer store hours, better after-sales service). Retailers will underprovide such services if they can be undercut by free-riding discount retailers who don’t provide these services. By eliminating retailer competition on prices for the product, it stimulates competition for better retail services. This rationale for RPM originates with a 1960 paper by Lester Telser of the University of Chicago.

This argument overlooks the possibility that different customers value these retail services differently. While the marginal customer is better off, some customers may place little value on such services, but they pay a higher price to cover retailers’ costs of providing them. As a result, fixed retail prices “may either improve or harm economic efficiency and welfare”. The harm may occur “because firms’ and consumers’ interests need not be aligned … [RPM] may well improve profits but hurt consumers and even reduce total welfare” (Jullien and Rey 2007, pp. 983-4).

Bruno Jullien and Patrick Rey also explain how RPM could facilitate collusion among producers. Because of the temptation to cheat on the collusive agreement, as explained in Mankiw’s text, producers would be better able to monitor each other’s behaviour with RPM because they can observe retail prices more easily than wholesale prices.

In short, depending upon circumstances, a retail price floor set by producers may improve or harm consumer welfare in a market. Mankiw is right in concluding that those in charge of enforcing antitrust laws don’t have an easy job in determining “what kinds of behavior actually impede competition and reduce economic well-being” (p. 349). However, his discussion of resale price maintenance is one-sided and incomplete.

 (ii) Could predatory pricing be rational?

Could it make sense for a firm to cut prices and incur losses to drive a competitor out of the market? Mankiw writes that “some economists are skeptical”, believing that this “it is rarely, if ever, a profitable business strategy”. He gives a detailed hypothetical example to illustrate the conclusion that “the predator suffers more than the prey”. However, the section finishes by acknowledging that economists continue to debate the issue and that “questions remain unresolved”, without giving any details (p. 300).

Given the weight that Mankiw places on the sceptics’ view, students would be surprised to know that “it is now the consensus view in modern economics that predatory pricing can be a successful and fully rational business strategy” (Bolton et al. 2000, p. 2242).

In a survey article on predatory pricing, Janusz Ordover contradicted Mankiw’s central story. He wrote, “the strategic approach to modelling pricing debunked the comfortable position that predation is more costly to predator than prey, and hence irrational and unlikely to occur” (2008, p.9).

What about evidence of how firms actually behave? “Recent empirical work has supported the rational predation models… thereby undermining the Chicago School’s claims about its irrationality” (Ordover 2008, p.4). This was based on both studies of legal cases, where firms had been prosecuted for predatory pricing, as well as historical case studies.

Pricing is not the only strategy a dominant firm (or firms) could use to drive out rivals or to deter their entry into the market. For example, in his review of antitrust policy, Oliver Williamson (1987, p.4) mentions the alleged predatory brand proliferation in the ready-to-eat cereals industry, investigated by the US Federal Trade Commission in the 1980s. The few firms that dominated the industry were suspected of deterring entry of competitors by differentiating similar products and increasing advertising spending. The largest four firms had more than 100 brands of cereals. With limited shelf space in supermarkets, brand proliferation and strong brand names helped to keep new entrants from the market (Morris 2019, p. 50).

Firms could also undertake pre-emptive investments, such as investment in excess production capacity. Such unused capacity could deter a potential entrant who would understand that its entry would be resisted with lower prices. The resulting increased demand could be met with the additional production capacity. The excess capacity could also be employed if price cuts were needed to discipline uncooperative rivals if oligopolistic firms were tacitly or overtly colluding.

(iii) Tying (and bundling)

In the section “Tying”, Mankiw gives an example of two movies which are offered as a package to movie theatres (p. 350). This is an example of what is termed pure bundling. If the theatres also had the option of buying each movie separately, that would be mixed bundling.

In a survey article, Barry Nalebuff describes tying as “a special case of mixed bundling; customers are offered prices for A and B together or for B alone, but not A without B” (2008, p.1). Typically, A and B are different products. Consider this example given by the US Federal Trade Commission (FTC).

The FTC challenged a drug maker that required patients to purchase its blood-monitoring services along with its medicine to treat schizophrenia. The drug maker was the only producer of the medicine, but there were many companies capable of providing blood-monitoring services to patients using the drug. The FTC claimed that tying the drug and the monitoring services together raised the price of that medical treatment and prevented independent providers from monitoring patients taking the drug. The drug maker settled the charges by agreeing not to prevent other companies from providing blood-monitoring services.

Mankiw’s example is not one of tying in the above sense, although the word ‘tying’ is used frequently in the 1962 Supreme Court decision about the bundling of movies that he refers to. His example is one of pure bundling of two similar products where the seller has a monopoly on both. The numerical example and the explanation for the bundling is adapted from Stigler (1963). Mankiw echoes George Stigler’s Chicago School argument that the monopolist is price discriminating to increase profits, but it can’t “increase its market power simply by bundling the two movies together” (p. 350).

Tying, as defined and illustrated above, does apply to Mankiw’s Microsoft example, where the Windows operating system was tied to the Internet Explorer browser. The two are different products and web browsers can be produced by a variety of firms.

Nalebuff writes that “the current literature suggests that, in a dynamic setting, bundling can profitably leverage market power by deterring entry, excluding one-good rivals, and amplifying existing market power” (2008, p.2). Mankiw essentially acknowledges this when he writes: “Yet economists have proposed more elaborate theories for how tying can impede competition”, but he gives no examples. Here is a simple one that illustrates how monopoly power in one market can increase market power in the market for the tied good.

[S]uppose that a restaurant in the only hotel on a resort island competes with local restaurants. If the hotel requires its guests to eat their meals at the hotel restaurant, then there may be fewer local restaurants as a consequence of the reduced patronage. Local residents will then have fewer alternatives, with the result that more of them may decide to frequent the hotel restaurant. In this case, tying can be profitable because it reduces competition in the tied market… (Carlton and Waldman, 2002, p. 195).

Mankiw concludes: “Given our current economic knowledge, it is unclear whether tying is adverse for society as a whole” (p. 350). However, it is clear that there are cases in which tying does reduce social welfare, potentially justifying regulatory action.

(iv) Summary

In each of the three examples of potentially anticompetitive behaviour, Mankiw provides a detailed description of the Chicago School position, which concludes that no regulatory action is necessary, while he briefly acknowledges that that is not the final word. In the conclusion of the chapter he writes that: “Although price-fixing among competing firms clearly reduces economic welfare and should be illegal, some business practices that appear to reduce competition may have legitimate if subtle purposes. As a result, policymakers need to be careful when they use the substantial powers of the antitrust laws to place limits on firm behavior.” (P. 353).

Readers can judge for themselves whether his unbalanced accounts leave the impression that antitrust authorities needn’t worry too much about strategies such as resale price maintenance, predatory pricing or tying.



Ballou, Brendan (2021) “The ‘No Collusion’ Rule, Stanford Law & Policy Review, 32: 213-52. Available here.

Bertrand, Joseph (1883) “Theorie Mathematique de la Richesse Sociale”, Journal des Savants, 67, pp. 499–508; translated by James W. Friedman in Andrew F. Daughety, ed., Cournot Oligopoly, Cambridge University Press, 1988, pp. 73–81.

Bolton, Patrick, Joseph F. Brodley and Michael H. Riordan (2000) “Predatory Pricing: Strategic Theory and Legal Policy”, Georgetown Law Journal, 88(8): 2241-2330.

Carlton, Dennis and Michael Waldman (2002) “The Strategic Use of Tying to Preserve and Create Market Power in Evolving Industries”, The RAND Journal of Economics, 33(2), pp. 194-220.

Cournot, Augustin (1838 [1960]) Researches Into the Mathematical Principles of the Theory of Wealth, English translation of 1897 reprinted by Augustus M. Kelly.

Ezrachi, Ariel and Maurice Stucke (2020) “Sustainable and unchallenged algorithmic tacit collusion”, Northwestern Journal of Technology and Intellectual Property, 17(2), 217-59. Available at:

Hovenkamp, Herbert J. and Fiona Scott Morton (2020) “Framing the Chicago School of Antitrust Analysis”, University of Pennsylvania Law Review, 168(7): 1843-78. Available at

Jullien, Bruno and Patrick Rey (2007) “Resale Price Maintenance and Collusion”, The RAND Journal of Economics, 38(4): 983-1001.

Levenstein, Margaret and Valerie Suslow (2008) “Cartels”, in Palgrave Macmillan (ed.), The New Palgrave Dictionary of Economics, DOI 10.1057/978-1-349-95121-5_1985-1.

Morris, P. Sean (2019) “Intellectual Property for Breakfast: Market Power and Informative Symbols in the Marketplace”, Cleveland State Law Review, 68(1): 36-72. Available at

Nalebuff, Barry (2008) “Bundling and Tying”, in Palgrave Macmillan (ed.), The New Palgrave Dictionary of Economics, DOI 10.1057/978-1-349-95121-5_2534-1.

O’Brien, Daniel and Greg Shaffer (1992) “Vertical control with bilateral contracts”, RAND Journal of Economics, 23(3): 299-308.

Ordover, Janusz (2008) “Predatory Pricing” in Palgrave Macmillan (ed.), The New Palgrave Dictionary of Economics, DOI 10.1057/978-1-349-95121-5_1778-2.

Stigler, George J. (1963) “United States v. Loew’s Inc.: A Note on Block-Booking”, The Supreme Court Review, Vol. 1963, pp. 152-157.

Telser, Lester G. (1960) “Why Should Manufacturers Want Fair Trade?” Journal of Law & Economics, Vol. 3, October, pp. 86-105.

Williamson, Oliver E. (1987) “Antitrust Policy”, in S. Durlauf, L.E. Blume (eds.), The New Palgrave Dictionary of Economics, DOI 10.1057/978-1-349-95121-5_680-1.


Related commentaries


Respond to this commentary

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

Please note that your email address will not be published.