Keeping Each Other in Check – Oligopolists' Strategic Interaction

George Orlov, Cornell University, Douglas McKee, Cornell University
Author Profile


Students will examine the behavior of two large oligopolists, deciding on which factors serve as the biggest deterrents to the firms' collusion. The main factor preventing collusion is the temptation to deviate from the agreement for both firms, which leads the Nash equilibrium in which both firms do not cooperate.

Context for Use

● The activity is appropriate for both Principles and Intermediate microeconomics, as well as an undergraduate Industrial Organization course.
● Ideally, students would need to have studied monopoly, oligopoly, and game theory. This exercise can be thought of in terms of the prisoners' dilemma or extrapolated from a Bertrand game.
● This activity should be suitable for discussion in any class size.
● The activity will likely take around 20 to 25 minutes.


In this activity students will examine the behavior of two large oligopolists, trying to decide which of the offered reasons is the most likely one behind the lack of collusion in the market. The main outcome of the exercise is that students will show that collusion by firms in a duopoly model is not sustainable/possible. The activity can also touch upon the use of Nash equilibrium in analyzing strategic firm interactions, if the timing is such that students are working on the exercise after covering Game Theory.

Applying both theoretic knowledge of oligopolistic markets and games, as well as real world facts, students will arrive at a conclusion that while there are legal collusion deterrents in place, firms' concern that their co-conspirator might deviate from the agreement and cause large losses for the firm that keeps to the agreement is the greatest deterrent to collusion among oligopolists.

Expected Student Learning Outcomes

In this exercise students will be able to evaluate different factors that prevent large firms, that are essentially oligopolists, from (tacit) collusion.

On the "big picture" scale, this relates to the goal of students recognizing real-world situations where the models they learn are (and are not) appropriate. Students should be able to apply their new tool set to simple real-world questions and use it to argue for (or against) particular policies.

Information Given to Students

Consider the fact that in the market for carbonated drinks in the US, the Coca-Cola Company and Pepsi Co. supply a tremendous share of the market (to be more specific, about 60% of it, split 40% and 20% respectively). This is especially true for cola-flavored drinks. Despite this, we do not see cola prices from Pepsi and Coca-Cola go much higher than other carbonated drinks (even if in the carbonated drinks market, and further, the cola subdivision of it, they do not have close substitutes). Why do the prices of Coke and Pepsi stay where they are, instead of substantially increasing? Discuss and arrange the following explanations in the order from most likely to least likely.

A. The government has strictly enforced anti-trust (i.e., anti-collusion) legislation which prevents the collusive price increase from occurring.
B. It is difficult for Coca-Cola and Pepsi to covertly negotiate a price increase.
C. Coca-Cola and Pepsi would be greatly concerned by the response of the smaller producers who might capture some of the market away from them despite brand loyalty.
D. Coca-Cola and Pepsi are concerned by each other's behavior even if they could covertly negotiate a collusive agreement and form a cartel.

student_handout_quotkeeping_each_ot.v2.docx (Microsoft Word 2007 (.docx) 17kB Sep16 19)

Teaching Notes and Tips

● The instructor should circulate around the classroom making sure that the groups are actively discussing the problem at hand. Since the activity calls for an ordered list of answers instead of a single choice by the teams, we suggest the following approach to reporting. First, discuss the most likely reason for collusion not taking place. Then, discuss the least likely reason. Finally, discuss the 2nd and 3rd choices in conjunction.

● Option D is the most likely reason for Coca-Cola and Pepsi not colluding on prices. Tacit collusion can be thought of as a Prisoner's Dilemma. In a context of a one-shot game, the payoff to deviating from colluding can be tempting enough for the game to have a unique non-collusion Nash Equilibrium (as in the standard Prisoner's Dilemma). It is a little trickier in a context of a repeating game, which is a suitable for Intermediate Microeconomics or introductory Industrial Organization and Game Theory courses. In a finite-horizon repeated Prisoner's Dilemma, the game unravels to the same equilibrium as a one-shot, since there does not exist an incentive not to deviate from collusion. Once the players know that deviating from collusion at time T is the best response, a backwards solution yields deviations at T-1, T-2, ... , 2, 1. In an infinitely played game (which some students might argue can act as a representation of the firms' strategic interactions - have them present supporting arguments why they think it is true), it is very likely that the discount factors do not yield a sufficiently large payoff to prevent a one time deviation from collusion in the first time period of the game. When talking about repeated games it is worthwhile to point out that the main issue is that the collusion equilibrium is not sustainable.

● Option B is the least likely deterrent to collusion, since having the right people play a game of golf together and arranging for safe communication channels is a very viable option.

● Options A and C are, arguably, equally likely. However, large companies have broken anti-trust laws before, maintaining successful cartels for years. The concern that the competitors producing substitutes to Coke and Pepsi might take a share of the market from the big two producers is a possibility. The complementing concern is consumers switching to healthier options is viable as well.

What kinds of follow-up questions are recommended for facilitating the debriefing conversation among team reporters? In particular, how might the instructor get teams to evaluate which answer is the best, provide the analytical support for the team answers, identify what information would enable an economist to decide between alternative answers? - Have the students think through how much impact on the price these firms would have and how much market power the competition/generics have, if they have not done so already.

What points should be emphasized in the instructor's summary remarks to conclude the exercise? - It would be good to talk about successful colluders not getting caught in the real world – all we know about cartels comes from the failed cartels and theoretic models of strategic interactions. Students should be informed that not all situations can be covered by this stylized model with stylized outcomes.


Reverse the direction of the question: ask the students to think of examples where they think the firm collusion (and illegal cartel formation) is prevented by antitrust legislation and where by the strategic environment they are in.


In 2014, a merger between Comcast and Time Warner Cable was prevented under the Sherman Act. Admittedly, the government failed to prevent the 2013 merger of the US Airways and American Airlines.

It is very likely that car manufacturers (for example, Toyota, Nissan, Mazda, and Honda) are prevented from further increasing their prices by the concern that even if they could arrange for a collusive price increase, the temptation to deviate and capture a higher share of the market would be to large.

References and Resources