CHAPTER 11 Imperfect Competition
11.1 What Does Equilibrium Mean in an Oligopoly?
11.2 Oligopoly with Identical Goods: Collusion and Cartels
11.3 Oligopoly with Identical Goods: Bertrand Competition
11.4 Oligopoly with Identical Goods: Cournot Competition
11.5 Oligopoly with Identical Goods but with a First-Mover: Stackelberg Competition
11.6 Oligopoly with Differentiated Goods: Bertrand Competition
11.7 Monopolistic Competition
11.8 Conclusion
In previous chapters, we studied the two ends of the market power spectrum: perfect competition and monopoly. In perfect
competition, a firm has no market power because it is only one of many producers in the market, the price is driven down to
marginal cost, and output is relatively high. In a monopoly, one firm has market power because it is the only producer of a good
in the market, price is greater than marginal cost, and output is lower. We also learned about the many pricing strategies that
firms with market power can use to earn greater economic profit.
Between these two ends of the spectrum are lots of industries that are neither perfectly competitive nor monopolistic. Coke and
Pepsi dominate the cola market together. Nintendo, Sony, and Microsoft dominate console video games. These companies
compete but are hardly perfectly competitive. Yet they aren’t stand-alone monopolies either. The industry structure between
perfect competition and monopoly is known as imperfect competition.
imperfect competition
The industry structure between perfect competition and monopoly.
This chapter introduces this important but sometimes complicated market structure. We begin by looking at several types of
oligopoly, a market structure characterized by competition among a small number of firms. Because there are many possible
ways in which oligopolistic firms compete, no single model of oligopoly exists that is applicable to every situation. Having a few
competitors in an industry — rather than many or only one — can lead to many possible price and output outcomes. It’s not as
simple as just picking the one where price equals marginal cost or the like. We might observe many outcomes, depending on the
market circumstances.
oligopoly
A market structure characterized by competition among a small number of firms.
With oligopolies, firms have some market power, but not necessarily monopoly power, and there is some competition, but not
perfect competition. We need to be a little more specific about aspects of the market before we can figure out what prices they
will charge, how much each company will produce, and how much profit each firm will earn. Just knowing how many companies
are in the market is not enough to know what will happen in an oligopolistic market. Industries with, say, four major firms can
look extremely different from each other. Other factors that have an effect on price and quantity decisions in an oligopoly
include: (1) whether the companies make identical products (as in an oil oligopoly) or products that are slightly different from
one another (like Coke and Pepsi); (2) how intensely the companies compete; and (3) whether they compete with one another by
choosing the prices they charge or the quantities they produce.
In this chapter, we present five of the most common models of how oligopolies behave, plus one additional model called
monopolistic competition, a type of imperfect competition where a large number of firms have some market power, but each
makes zero economic profit in the long run. Whenever you have this many models as possible explanations for market behavior,
it’s important to determine which one is appropriate for a specific case. This decision isn’t always obvious in practice, so we
discuss some ideas for determining which model is most appropriate for various real-world situations.
monopolistic competition
A type of imperfect competition where a large number of firms have some market power, but each makes zero economic profit in the long run.
11.1 What Does Equilibrium Mean in an Oligopoly?
Before we introduce the different models of oligopoly, we need to lay some groundwork. Specifically, we have to expand on the
idea of what an equilibrium is for these industries. The concept of equilibrium in perfect competition and in monopoly is easy. It
means a price at which the quantity of the good demanded by consumers equals the quantity of the good supplied by producers.
That is, the market “clears.” The market is stable with no excess supply or demand, and the consumers and producers do not want
to change their decisions.
The problem with applying this idea of equilibrium to an oligopolistic industry is that each company’s action influences what the
other companies want to do. To achieve an outcome in which no firm wants to change its decision means determining more than
just a price and quantity for the industry as a whole. It has to apply to each firm individually, too.
An equilibrium in an oligopoly starts with the same idea as in perfect competition or monopoly: The market clears. However, it
adds a requirement that no company wants to change its behavior (its own price or quantity) once it knows what other companies
are doing. In other words, each company must be doing as well as it can conditional on what other companies are doing.
Oligopoly equilibrium has to be stable not only in equating the total quantities supplied and demanded, but also must be stable
among the individual producers in the market.
An equilibrium in which each firm is doing its best conditional on the actions taken by other firms is called a Nash equilibrium.
It is named after Nobel laureate John Nash (who was also the subject of the award-winning book and movie titled A Beautiful
Mind). The Nash equilibrium concept is even more central in the next chapter when we study game theory, which explores the
strategic interaction among firms. For our purposes in this chapter, though, the following example will help clarify what is and
what is not a Nash equilibrium in an oligopoly.
Nash equilibrium
An equilibrium in which each firm is doing its best conditional on the actions taken by other firms.
Application: An Example of Nash Equilibrium: Marketing Movies
Major superhero action movies like Disney’s Black Panther or Warner Brothers’ Wonder Woman are amazingly expensive to make. By the time the
studios have paid for the CGI, the actors, and everything else, they’re looking at a bill of around $180 million.1 But on top of these production costs,
Disney and Warner Brothers each then had to pay many more millions of dollars on advertising and marketing the films so people would watch their
movies.
Let’s suppose Disney and Warner Brothers are the only two movie companies that make superhero feature films, and that their advertising influences
people’s choices of what movie to see. Advertising doesn’t increase the overall number of movies people watch, just which movie they do.
Now think of both studios planning to release the next installments in these series, Black Panther 2 and Wonder Woman 2, on the same summer
weekend. Furthermore, let’s assume that the cost of production is still $180 million and the cost of advertising is $70 million. If both studios
advertise and compete with each other, their marketing efforts will cancel out. As a result, the two will split the market, and each will bring in, let’s
say, $500 million of revenue. Subtracting the $180 million production cost and the $70 million advertising cost, that leaves $250 million of profit to
each studio.
If, on the other hand, the studios could somehow agree not to advertise at all, they would again split the market, but this time each would save the
$70 million in advertising costs. In this case, the studio profits would be greater at $320 million each.
Disney and Warner Brothers would prefer the second, higher-profit outcome. The problem is that, due to the nature of advertising’s influence on
moviegoers, if only one studio advertises and the other doesn’t, then the studio that advertises will get a larger share of the audience and the other one
will be left with less. Suppose, for example, that the studio engaged in advertising would earn $800 million of revenue, and the other would earn only
$100 million. The firm advertising its film therefore earns a profit of $550 million ($800 million of revenue minus the $180 million production cost
11.2 Oligopoly with Identical Goods: Collusion and Cartels
Model Assumptions
Collusion and Cartels
Firms make identical products.
Industry firms agree to coordinate their quantity and pricing decisions, and no firm deviates from the agreement even if breaking it is in the firm’s best
self-interest.
In the next several sections, we examine several different models of imperfect competition. They give very different answers about the
way in which firms make decisions, so it’s important to know which model is the right one to use. A box at the start of each section lists
the conditions an industry must meet for that model to apply. In the first model, all the firms in an oligopoly coordinate their production
and pricing decisions to collectively act as a monopoly to gain monopoly profits to be split among themselves. This economic behavior is
known as collusion. The organization formed when firms collude is often called a cartel.2
collusion
Economic behavior in which all the firms in an oligopoly coordinate their production and pricing decisions to collectively act as a monopoly to gain monopoly profits
to be split among themselves.
cartel
The organization formed when firms collude.
If the companies in an oligopoly can successfully collude, figuring out the oligopoly equilibrium is easy. The firms act collectively as a
single monopolist would, and the industry equilibrium is the monopoly equilibrium (output is the level for which
and the price is determined by the demand curve, as we saw in Chapter 9).3 Don’t try this at home,
though. Cartels and collusion violate the law in most every country of the world, and in the United States, it is a criminal offense that has
landed many executives in prison. We discussed in Chapter 9 that governments enforce antitrust laws because of monopolies’ potential to
harm consumers. That explains why collusion has to be done in secret. Interestingly, the secrecy itself makes it more difficult for cartels
to maintain a stable equilibrium.
FREAKONOMICS
Apple Always Wins, or Does It?
In the months leading up to the launch of the iPad, Apple was also preparing to open the iBooks Store, which would allow users to buy and read books on
mobile devices. But Apple faced a dilemma. It would be difficult for the iBooks Store to succeed unless it could match the $9.99 price point of Amazon’s
Kindle, its major competitor. This price, thought by some to be less than marginal cost, had served Amazon well in building a customer base for e-books, but
the price made it challenging for a new entrant like Apple to successfully compete.
Amazon’s low price point was a concern for book publishers as well. They feared that cheap e-books would hurt sales of their more expensive print copies and,
over the longer term, influence the public’s expectations regarding book prices. Amazon’s growing market power also posed the threat that Amazon might start
directly competing with publishers.
The major publishers had already begun engaging in talks, meeting in private dining rooms in New York City restaurants to discuss ways to force Amazon to
price above $9 99. Before one of their meetings, David Young, then chairman and CEO of Hachette Book Group, told a fellow publisher, “I hate [Amazon’s]
bullying behavior and will be happy to support a strategy that restricts their plans for world domination.” The publishers started implementing their strategy.
They coordinated on raising the wholesale price of e-books and introducing “windowing,” which delayed e-book versions of new releases to protect hardcover
sales. After one particular correspondence regarding these tactics, Young advised another publishing executive that “it would be prudent for you to double
delete this from your email files.”
Eddy Cue, Apple’s Senior Vice President of Internet Software and Services, learned of the publishers’ discontent with Amazon’s price point, and he began
requesting meetings with the major publishers to encourage them to join the iBooks Store. Cue assured the publishers that Apple would price books higher than
Amazon. After the first of his meetings, he reported to Apple’s then-CEO Steve Jobs that the publishers were “ecstatic” about the prospect of Apple’s entry into
the industry.
Over the next several weeks, the publishers and Apple crafted a contract that featured a market-wide transition from a wholesale model to an agency model
(whereby the publisher rather than the retailer sets the retail price) and a clause that guaranteed Apple would hold the lowest prices on the market. These
changes would effectively drive e-book prices upward while securing Apple’s competitive position in the market.
Amazon responded by also moving to an agency model in the following months. Soon thereafter, e-book retail prices increased an average of 14.2% per unit
and 42.7% for New York Times bestsellers. Publishers sold fewer e-books through Amazon, as might have been expected, though estimates of the magnitude of
this decrease varied. On the whole, however, the collusion between Apple and the publishers appeared to be a massive success. Not only had it raised prices, it
seriously eroded Amazon’s monopoly. Before the price-fixing scheme, Amazon held about 90% of the e-book market. A year-and-a-half after the kickoff of the
iBooks Store, when the scheme was in full effect, this share was closer to 60%.
Once again, it seemed as if the old Apple magic had worked: The company revolutionized yet another market by entering it. Maybe that would have been the
case had things stayed the way they were. But any cheers of victory among Apple and the publishers quickly faded when the U.S. government sued them for
collusion. The presiding judge announced her decision just over a year later: Apple was guilty. The words the judge used were scathing. “To adopt Apple’s
theory, a fact-finder would be confronted with the herculean task of explaining away reams of documents and blinking at the obvious.” Several states and
private plaintiffs sought more than $800 million in damages. Apple managed to soften the blow by appealing and reaching a $450 million settlement in case its
appeal fell through, but it lost the most important gain the scheme had rendered: its ability to control e-book prices.
The Instability of Collusion and Cartels
The firms in an oligopoly would love to collude. They could earn more profit. Adam Smith, the eighteenth-century philosophy professor
and one of the fathers of the discipline of economics, recognized this. He wrote in The Wealth of Nations, “People of the same trade
seldom meet together, even for merriment and diversion, but the conversation ends in a conspiracy against the public, or in some
contrivance to raise prices.”
But colluding is harder than it looks. Each member of a cartel has an incentive not to go along. Although firms in a market might be able
to come to some initial agreement over a bargaining table, collusion turns out to be unstable — not an equilibrium.
Think about an industry in which there are two firms, Firm A and Firm B, that want to collude. To keep things simple, say both firms
have the same constant marginal cost c. If the two firms act collectively as a monopolist, we can follow the monopoly method from
Chapter 9 to determine the market equilibrium. Each firm will operate where marginal revenue equals marginal cost. It’s not stable,
though, because each will want to increase its output at the other’s expense.
Suppose the inverse market demand curve for their product is , where P is the price per unit and Q
is the quantity produced. We know from Section 9.2 that the marginal revenue curve corresponding to this linear inverse demand curve is
. The firms will produce a quantity that sets their marginal revenue equal to their
marginal cost c:
Solving this equation for Q gives . This is the industry’s output when its firms collude to
act like a monopolist. If we plug this back into the demand curve equation, we find the market price at this quantity:
.
The advent of artificial intelligence (AI) and other algorithmic business practices, however, has recently raised another concern. What if collusion isn’t the
result of people actually deciding to act in a particular way, but rather because pricing algorithms built into software “decide” to raise prices? Suppose
competing companies wrote pricing software that effectively operationalized the command, “If our competitors raise prices, raise our price.” It isn’t difficult to
imagine that the market could fall into a cartel-like outcome without any humans agreeing with one another to actually take the step of increasing prices. How
could antitrust authorities prosecute a case where people never made a collusive deal? You can’t put an algorithm in jail.
Perhaps we could make even using such cartel-spurring algorithms illegal and subject the installers to penalties. But a while a strategy to “price high if they
price high” is transparent and easy to track, most pricing algorithms are much more complex. They could plausibly lead to cartel-like outcomes in ways that no
one, possibly even their creators, might have imagined. Several AI bots given no other guidance than to set prices to maximize profits might quickly discover
for themselves that the best way to do this is to collude. (They would also be fairly good at quickly detecting any cheating behavior, raising the stability of the
collusive outcome.)
Policymakers and economists have yet to settle on a recommended course of action. But technological trends make it likely that these sorts of situations will
occur with increasing frequency in the future.
11.3 Oligopoly with Identical Goods: Bertrand Competition
Model Assumptions
Bertrand Competition with Identical Goods
Firms sell identical products.
The firms compete by choosing the price at which they sell their products.
The firms set their prices simultaneously.
In the previous section, we learned that the collusion/cartel model of oligopoly in which firms behave like a monopoly is unlikely
to hold in reality because coordination is not an equilibrium and the agreement will likely break down. We need a model in which
firms compete against one another. The first such model is as simple as it gets: Firms sell the same product, and consumers
compare prices and buy the product with the lowest price. This oligopoly model in which each firm chooses the price of its
product is called Bertrand competition, after Joseph Bertrand, the nineteenth-century French mathematician and economist who
first wrote about it. When firms are selling identical products, as we’re assuming here, Bertrand oligopoly has a particularly
simple equilibrium: , just like perfect competition. In later sections of this chapter, we see how
circumstances change when firms sell products that are not identical.
Bertrand competition
Oligopoly model in which each firm chooses the price of its product.
Setting Up the Bertrand Model
To set up this model, let’s suppose a market with only two companies in it exists. They sell the same product and have the same
marginal cost. For example, suppose there are only two stores in a city, a Walmart and a Target, and these stores are located next
to each other. They both sell the latest Nintendo Switch and each firm’s marginal cost is $300 per console. This includes the
wholesale price the firm has to pay Nintendo as well as miscellaneous selling costs, such as stocking the consoles on shelves,
checking customers out, and so on.
We need one further assumption: Consumers don’t view either store differently in terms of service, atmosphere, or the like. If
consumers value these characteristics separately from the consoles, then in a way the products would no longer be identical and
we would need to model the firms’ behavior using the model of differentiated products discussed later in the chapter.
With only two companies in a market, it might seem as if there would be a lot of market power and high markups over cost. But
suppose the customers in this market have a simple demand rule: Buy the console from the store that sells it at the lowest price. If
both stores charge the same price, consumers flip a coin to determine where they buy. This rule means, in effect, that the store
charging the lowest price will garner all the demand in the market. If both stores charge the same price, each store gets half the
demand.
Suppose the total demand in the market is for Q consoles. Let’s denote Walmart’s price as and Target’s price as
. The two stores then face the following demand curves:
Demand for Nintendo Switches at Walmart:
Demand for Nintendo Switches at Target:
Each store chooses its price to maximize its profit, realizing that it will sell the number of units according to the demand curves
above. We’ve assumed the total number of consoles sold, Q, doesn’t depend on the price charged. The price only affects which
store people buy from. (We could alternatively have allowed Q to depend on the lowest price charged; all the key results
discussed below would remain the same.)
Nash Equilibrium of a Bertrand Oligopoly
Remember that in a Nash equilibrium, each firm is doing the best it can given whatever the other firm is doing. So to find the
equilibrium of this Bertrand model, let’s first think about Target’s best response to Walmart’s actions. (We could do this in the
reverse order if we wanted.) If Target believes Walmart will charge a price for Nintendo Switches, Target will sell
nothing if it sets its price above , so we can probably rule that out as a profit-maximizing strategy. Target is left with
two options: Match Walmart’s price and sell units, or undercut Walmart and sell Q. Because all it has to do is
undercut Walmart by any amount, dropping its price just below will only reduce its per-unit margin by a tiny amount,
but the store will double its sales because it will take the whole market instead of splitting it.
As an example, suppose and Target thinks Walmart will charge
. If Target also charges , it will sell 500 consoles at a
profit of $25 each (the $325 price minus the $300 marginal cost). That’s a total profit of $12,500. But if Target charges $324.99,
it will sell 1,000 Nintendo Switches at a profit of $24.99 each. This is a profit of $24,990 — almost double what it was at $175.
Target has a strong incentive to undercut Walmart’s expected price.
Of course, things are the same from Walmart’s perspective: It has the same incentive to undercut whatever price it thinks Target
will choose. If it believes Target is going to charge for a Nintendo Switch, Walmart
could price its consoles at $324.98 and gain back the entire market. But then Target would have the incentive to undercut this
expected price, and so on.
This incentive for undercutting would only stop once the price each store expects the other to charge falls to the level of the
stores’ marginal costs ($300). At that point, cutting prices further would let a store gain the entire market, but that store would be
selling every Switch at a loss.
The equilibrium of this Bertrand oligopoly occurs when each store charges a price equal to its marginal cost — $300 in this
example. Each obtains half of the market share, and each store earns zero economic profit. The stores would like to charge more,
but if either firm raises its price above marginal cost by even the smallest amount, the other firm has a strong incentive to
undercut it. And dropping prices below marginal cost would only cause the stores to suffer losses. Thus, the outcome isn’t great
for the firms, but neither firm can do better by unilaterally changing its price. This is the definition of a Nash equilibrium.
In the identical-good Bertrand oligopoly, one firm cannot increase its profit by raising its price if the other firm still charges a
price equal to its marginal cost. If the firms could somehow figure out a way to coordinate changes in their actions so that they
both raised prices together, they would raise their profits. However, the problem with this strategy, as we saw earlier, is that
collusion is unstable. Once the firms are charging prices above marginal cost, a firm can raise its profits by unilaterally changing
its action and lowering its price just slightly.
The Bertrand model of oligopoly shows that even with a small number of firms, competition can still be extremely intense under
the right conditions. In fact, the market outcome of Bertrand competition with identical goods is the same as that in a perfectly
competitive market: Price equals marginal cost. This super-competitiveness occurs because either firm can steal the whole
market away from the other by dropping price only slightly. The strong incentive to undercut the price leads both firms to drop
their prices to marginal cost.
This example had only two firms, but the result would be the same if there were more. The intuition is the same: Every firm’s
price-cutting motive is so strong that the only equilibrium is for them to all charge a price equal to marginal cost and split the
market evenly.4 The strong assumptions of the Bertrand model with identical products are rare, but some online markets
approximate this condition. Where comparing across merchants is really easy, the lowest-priced seller can take the lion’s share of
the market and these markets often end up with all firms charging the same low price, as the model predicts.