L09 Oligopoly
L09 Oligopoly
Learning Outcomes
By the end of this chapter, and having completed the Essential reading and activities, you should
be able to:
Define the conditions for various Oligopoly model.
Understand the application of game theory into various game theoretic Oligopoly models.
Define the market equilibrium price and quantity under various Oligopoly models.
Describe the society welfare under various Oligopoly models.
Essential Reading
Morgan, Katz and Rosen (MKZ) chapter 15: Oligopoly and strategic behaviour
Perloff (P) chapter 14: Oligopoly
Pindyck and Rubenfeld (P&R) chapter 12: Monopolistic competition and oligopoly;
chapter 13: Game Theory and Competitive Strategy
The main difference between Oligopoly and other markets lie in strategic behaviour. In
monopolistic competition, there are a few firms in the market but the firms do not
consider their opponents’ strategies.
In Oligopoly competition, there are also a few firms and but they consider their rivals’
strategies prior to their decision making. For example, how much to produce, how much
is the price of their good, which firm moves first, is it a one-time game or repeated game
etc.?
Oligopoly
Repeated
Non-Cooperative Cooperative
games
Important definitions
Price competition: Firms compete with each other by choosing the price for their
product.
Bertrand model: Firms produce the same good and choose the price simultaneously.
Cournot model: Firms produce the same good and choose their output
simultaneously.
Stackelberg model: Firms produce the same good, but there is sequential quantity
competition among firms, where one of the firm chooses its output first and the rest
firms will learn the leader’s production decision and decide on their productions.
Here, we will examine a one-period quantity competition model where each firm has to
guess the other firm’s output choice, and then the firm will produce its own profit-
maximising output based on the guess.
Note that firms will guess each other’s output level but they don’t know the exact output
level! Assuming there are only 2 firms in the Cournot model and, for simplicity, suppose
both Firm 1 and 2 have similar cost functions. A description of the model is shown
below:
Algebraic example
For Firm 1,
7) Market quantity,
8) Market price,
CS
E
PS DW
c MC
D Q
Q2
R2
R1
Q1
The Cournot-Nash equilibrium E is simply the pair of outputs produced by both firms
where the two reaction function curves intersect. At that point, both firms are producing
a profit-maximising level of output given the output choice of their rival.
As one can see, if Firm 1 and 2 have similar cost functions and are facing the same
market demand, one only has to work out one of the firm’s reaction functions like that of
Firm 1’s reaction function. Firm 2’s reaction function is the just the mirror of Firm 1’s
reaction function.
For Firm 1,
5. By symmetry, due to similar cost functions, Firm 2’s reaction function is,
7. Market quantity, 6
8. Market price, 10 6 4
10
CS
E
4
PS DW
1 MC
D Q
6 10
Q2
R2
4.5
E
3
R1
Q1
3 4.5
The Cournot-Nash equilibrium is E, where Firm 1 and 2 each produces 3 units of output.
10
Monopoly
5.5
4 Cournot
Perfect Competition
1 MC
D Q
4.5 6 10
For Firm 1,
1. Firm 1’s profit function, 40 1
7. Market quantity, 26
8. Market price, 40 26 14
4
0
CS
E
1
4
1
0
PS
DW
1 MC
D D’ Q
1 2 4
0 6 0
A shift in the demand curve will increase the firms’ output and profit, market output and
price.
For Firm 1,
7. Market quantity, 4
8. Market price, 10 4 6
10
CS
E
6
PS DW
4 MC’
1 MC
D
Q
4 10
Q2
R2
E1
3
2 E2
R1
Q1
2 3
An increase in the cost function will decrease the firms’ output and profit, market output
and price. The reaction function curves of the firms will shift inward too.
For Firm 1,
1. Firm 1’s profit function, 10 1
9 2 0
For Firm 2,
1. Firm 2’s profit function, 10 4
2. For profit-maximisation,Max 6
Market quantity, 5
Market price, 10 5 5
Firm 1’s profit, 5 1 4 16
Firm 2’s profit, 5 4 1 1
10
CS
E
5
PS DW
1 MC
D Q
5 10
Q2
R2
E
1
R1
Q1
4 4.5
Due to asymmetric cost, the output and profit of Firm 1 and 2 will be different.
For Firm 1,
1. Firm 1's profit function
Becomes ∑
Becomes ∑
2 0
∑
4. Obtain the reaction function of firm 1,
5. By symmetry, due to similar cost functions, firm i’s reaction function is:
1
2
7. Market quantity, ∑
8. Market price,
You can plug in the number of firms to check the validity of this equation. Try pluggin in
N = 2 and obtain the same equations from the earlier example. In the exam, there is a
possibility that the examiner will ask you to obtain the number of firms when the
equilibrium profit is zero. You can solve it by simply setting 0, and find
out the value of N.
For Firm 1,
1. Firm 1’s profit function, 9
7. Market quantity, 6
8. Market price, 10 6 4
Comparing to the earlier example that has a cost function without fixed cost, the
presence of fixed cost does not affect the firm’s reaction function and output decisions;
the market equilibrium quantity and price. However, it affects the firm’s profit. Here, if
the fixed cost, F > 9, Firm 1 will make losses and will exit the market.
For Firm 1,
Hence, Firm 1’s output 2 and then substitute 2 into Firm 2’s reaction
function to solve for Firm 2’s output, for which Firm 2’s output 3.5
Q2
R2
Production cap
4.5
3.5 E
3
R1
Q1
2 3 4.5
Here, we will examine a one-period (2-stage) quantity competition where one firm (also
known as leader) has the first-mover advantage to choose its output first, and then
another firm (the follower) will produce its own profit-maximising output with reference
to the leader’s output.
Note that when the follower makes its decision, it already knows the leader’s output
decision! Assuming there are only 2 firms in the Stackelberg model and, for simplicity,
suppose both Firm 1 and 2 have similar cost functions, a description of the model is
shown below:
Example
Assuming Firm 1 moves first and chooses its output. Subsequently, Firm 2 chooses its
profit-maximising output with reference to Firm 1’s output. Since Firm 1 has the first
mover advantage, it will maximise its profit by including Firm 2’s reaction function into its
own profit function.
Since the interaction is a sequential game with 2 players, the steps for solving the
Stackelberg equilibrium is based on the concept of backward induction. We will start
with the follower, Firm 2:
1. Firm 2’s profit function,
2 0
5. Substitute Firm 2’s reaction function into Firm 1’s profit function,
2 2
9. Market quantity,
16
CS
E
PS DW
c MC
D Q
Q2
R2
Cournot
Stackelberg
R1
Q1
The Stackelberg equilibrium S, is simply the pair of outputs of both firms where Firm 1
produces its output first after guessing the output decision of Firm 2. At that point, both
firms are producing a profit-maximising level of output given the output choice of their
rival.
Since the interaction is a sequential game with 2 players, the steps for solving the
Stackelberg equilibrium is based on the concept of backward induction. We will start
with the follower, Firm 2:
5. Substitute Firm 2’s reaction function into Firm 1’s profit function,
9
9
2
4.5
2
4.5
Under symmetric cost, follower output is always
2.25 half of the leader’s output.
10
CS
E
3.25
PS DW
1 MC
D Q
6.75 10
Q2
R2
Cournot
3
Stackelberg
2.25
R1
Q1
3 4.5
The Stackelberg equilibrium is S, where Firm 1 produces 4.5 units of output and Firm 2
produces 2.25 units of output.
The consumer surplus, 10 3.25 6.75 22.78
The industrial profit / producer surplus, 3.25 1 6.75 15.19
The deadweight loss, 3.25 1 9 6.75 2.53
The following graph shows a simple comparison among Cournot, monopoly and perfect
competition:
10
Monopoly
5.5
4 Cournot
Stackelberg
3.25
Perfect Competition
1 MC
D Q
4.5 6 6.75 9 10
Perfect
Monopoly Cournot Stackelberg
Competition
Consumer
10.125 18 22.78 40.5
Surplus
Producer Surplus 20.25 18 15.19 0
Deadweight Loss 10.125 6 2.53 0
Since the interaction is a sequential game with 2 players, the steps for solving the
Stackelberg equilibrium is based on the concept of backward induction. We will start
with the follower, Firm 2:
5. Substitute Firm 2’s reaction function into Firm 1’s profit function,
39
39
2
19.5
2
19.5
40
CS
E
10.75
10
PS
DW
1 MC
D D Q
10 29.25 40
A shift in the demand curve will increase the firms’ output and profit, market output and
price.
Since the interaction is a sequential game with 2 players, the steps for solving the
Stackelberg equilibrium is based on the concept of backward induction. We will start
with the follower, Firm 2:
5. Substitute Firm 2’s reaction function into Firm 1’s profit function,
6
6
2
3
2
10
CS
E
5.5
PS DW
4 MC’
1 MC
D Q
4.5 10
Q2
R2
Stackelberg
2.25
R1
1.5
Q1
3 4.5
An increase in the cost function will decrease the firms’ output and profit, market output
and price. The reaction function curve of firms will shift inward too.
Since the interaction is a sequential game with 2 players, the steps for solving the
Stackelberg equilibrium is based on the concept of backward induction. We will start
with the follower, Firm 2:
For Firm 1,
6. Substitute the Firm 2’s reaction function into Firm 1’s profit function,
8
9 5
2 2
10
CS
E
3.5
PS DW
1 MC
D Q
6.5 10
Q2
R2
Stackelberg
1.5
R1
Q1
5
Due to asymmetric cost will, the output and profit of firm 1 and 2 will be different.
Since the interaction is a sequential game with 2 players, the steps for solving the
Stackelberg equilibrium is based on the concept of backward induction. We will start
with the follower, Firm 2:
5. Substitute Firm 2’s reaction function into Firm 1’s profit function,
9
9
2
4.5
2
4.5
10.125
5.0625
In a similar fashion to the Cournot model, the presence of fixed cost does not affect the
firms’ reaction function, the market equilibrium quantity and price. However, it will affect
the firms’ profit and thus the firms’ decision to stay or exit from the market.
For example, if 5.0625, which implies Firm 2 is making losses, then Firm 2 will exit
the market and Firm 1 becomes the monopolist in the market. This will be discussed
further during a later part - Entry Deterrence and Accommodation.
For Firm 3,
1. Firm 3’s profit function,
10 1
For Firm 2,
6. Substitute the Firm 3’s reaction function into Firm 2’s profit function,
9 1 1
9 4.5
2 2 2
For Firm 1,
11. Substitute the reaction functions of Firms 2 and 3 into Firm 1’s profit function,
9 q 9 q q
π 9 q q
2 2
9 q 9 q
π 9 q
2 4
9 q
π q
4 4
17. Optional Material: One Leader and N Followers in the Cournot Model
This is a more complicated case that combines Stackelberg and Cournot models. For
simplification and easier understanding, we will work with real numbers. Assuming there
are 3 firms: Firm 1, 2 and 3 in a market. Firm 1 has first-mover advantage and decides
its output first; subsequently Firm 2 and Firm 3 observes Firm 1’s output and produce
their output simultaneously. Note that Firm 2 and 3 are engaged in Cournot
competition.
First, we adopt the backward induction strategy and solve the Cournot equilibrium.
Now, after obtaining the output of Firm 2 and 3, we can substitute it into Firm 1’s profit
function and work out the output of Firm 1 in the Stackelberg model.
For Firm 1,
7. Firm 1’s profit function,
10 1
8. Substitute the reaction function of Firm 2 and 3 into Firm 1’s profit function,
9 9
9 3
3 3 3
Assuming firms are producing a homogenous good and there is symmetric information,
under simultaneous price competition, consumers will buy the good from the firm that
offers the lowest price.
Further assuming firms do not have capacity constraints and have production that
exhibits constant returns-to-scale, logically, firms will have the incentive to offer a price
that is lower than what other firms are selling in order to capture the entire market
demand.
However, the lowest price that the firm can offer is its marginal cost. Consequently, all
firms will set its price such that . If all firms have the same cost function, all
firms will sell their product at the same price and share the market equally. If one firm
has lower MC than others, then this firm will set a price equal to the second lowest MC
minus a small amount .
This is the equilibrium because there is no incentive for firms to increase or decrease
the price. If Firm 1 increases its price such as , it will lose its market share and
Firm 2 will capture the entire market demand. If Firm 1 decreases its price such that
, it will capture the entire market but it will also make losses because
.
However, the Bertrand model does not imply perfect competition, although perfect
competition can be considered as one form of the Bertrand model! Consider the
following model:
Assuming inverse market demand, 10 , ∑
Since Firm 1 has 2, Firm 2 could capture the entire market by pricing its output
at 2, such as 1.99. At the Bertrand equilibrium, Firm 2 makes
positive profit which violates the perfect competition model.
Meanwhile, this is the Bertrand equilibrium because there is no incentive for Firm 2 to
change its price. If Firm 2 increase the price, 2, Firm 1 will capture the entire
market demand at a lower price. If Firm 2 decreases the price to 2, its profit will
fall.
Besides, Firm 2 will be the monopoly in the market because Firm 1 is unable to compete
with Firm 2 in terms of price and will leave the industry.
This example also shows that it is not necessary that a monopolist will always price its
goods at . With only one firm in the market, there are many ways for the
monopolist to set the price for its output. The following ways have been covered:
19. Collusion
All three models we learnt earlier, namely Cournot, Stackelberg and Bertrand are non-
cooperative and one-period games.
Instead of competing with each other in the form of quantity and price, firms might
choose to cooperate and form a monopoly like a cartel. In order to form a monopoly in
the market, the member firms must collectively decide:
Under cooperation, both firms form a cartel and act like a monopolist. Assuming the
collusion is such that firms:
Amidst the interaction between Firm 1 and 2, there are four possible scenarios in terms
of payoff for firms.
Firm 2
Cooperate Cheat
Cooperate Cooperate, Cooperate Cooperate, Cheat
Firm 1
Cheat Cheat, Cooperate Cheat, Cheat
10
E
5.5
1 MC
D Q
4.5 9 10
The steps to obtain the equilibrium when one firm cheats and another firm cooperates.
The outcome when firm 2 cooperates while firm 1 takes advantage of firm 2’s
cooperative behaviour is shown in the following graph:
Q2
R2
Collusion
2.25
R1
Q1
3.375
10
E
4.375
1 MC
D Q
5.625 9 10
7. Market quantity, 6
8. Market price, 10 6 4
The following graphs show the outcome when both firms refuse to cooperate.
Q2
R2
4.5
Cournot
3
One firm cheating
2.25
Collusion
R1
Q1
3 3.375
10
E
4
1 MC
D Q
6 9 10
Firm 2
Cooperate Cheat
Cooperate 10.125, 10.125 7.59, 11.39
Firm 1
Cheat 11.39, 7.59 9, 9
Obviously, the dominant strategy for both firms is to cheat and the Nash equilibrium is
(9, 9). Notice that it is a prisoner’s dilemma game where both firms could have earned
more profit if they cooperated, but they will cheat because it is their dominant strategy.
Even if it is not a one-period game, as long as it is finite game, it will still yield the same
outcome in that both firms will cheat by the logic of backward induction.
Assuming it is a three-period game, the dominant strategy for both firms in the last game
is to cheat. Since both firms realise that the other player will cheat in the last game,
they will cheat in the second last game, and then the second last game too. In the end,
if this is a finite game, both players will cheat in every period.
In short, if the competition is repeated over a finite number of periods, firms will play
according to the Nash equilibrium of the one-period game in each period and there is no
room for collusion.
However, if the game is played for infinite periods, then there is a possibility that firms
might collude instead of engaging in competition whether in terms of price or quantity.
One of a strategies to elicit collusion is the trigger strategy. Specifically, the strategy
stipulates that players will cooperate so long as the other players also cooperate. They
will switch to compete for the remaining periods if one of the players are found out to be
cheating.
Cheating gives highest level of profit. This is followed by the level of profit if both firms
cooperate. When both firms compete, the level of profit is the lowest. The trigger
strategy thus induces cooperation as it does not pay to cheat and earn a one-off
attractively high profit and then the lowest amount of profits for all subsequent periods.
Knowing that Firm 2 could enter the market, Firm 1 could deter Firm 2’s entry if it pays.
If not, Firm 1 will accommodate the entry. The structure of interaction is shown in the
following decision tree.
Payoff
Firm 1 Firm 2
Deter 10 0
entry
Incumbent
Firm 1
Enter
Sub‐game
Entrant Accomodate
entry 10.125 1.0625 Perfect Nash
Firm 2 Equilibrium
No Enter Monopoly
Incumbent 20.25 0
Firm 1
First scenario: firm 2 stays out and firm 1 becomes the monopoly in the market.
Since Firm 1 decides its output in Period 1 and Firm 2 decides its output in Period 2,
Firm 1 has first-mover advantage, it would be a Stackelberg equilibrium.
Since the interaction is a sequential game with 2 players, the steps for solving the
Stackelberg equilibrium is based on the concept of backward induction. We will start
with the follower, Firm 2:
5. Substitute Firm 2’s reaction function into Firm 1’s profit function,
9
9 4.5
2 2
10.125
5.0625
First, we need to think: under what kind of situation will Firm 2 not enter the market?
Answer: Firm 2 will not enter the market if it makes losses or zero profit.
3. Substitute the reaction function of Firm 2 into its profit function, Firm 2’s profit
function becomes,
9 9
9
2 2
9 9
9
2 2
9
2
4. Set Firm 2’s profit to zero to find out the value of q1 that will cause Firm 2 to earn
zero profit. When Firm 2’s profit equals zero:
9
0
2
5. Firm 1’s output,
9 2√
7. As then 2 if 5.
Market output, 7
Market price, 10 3
Firm 1’s profit, 3 1 5 10
Firm 2’s profit, 9 9 2 5 2 4 0
Notice that Firm 1’s profit is higher if it accommodates than if it deters Firm 2. The sub-
game perfect equilibrium is then (10.125, 1.0625) that can be obtained through the
backward-induction method.
Specifically, if Firm 2 enters, Firm 1 will choose to accommodate entry as profit from
accommodating > profit from deterring, i.e., 10.125 > 10. Secondly, Firm 2 knows that
Firm 1 will accommodate its entry, it will decide to enter as entering gives it higher profit
than staying out, i.e., 1.0625 > 0
What if F = 9?
Firm 1 output, 9 2√ → 3
It implies Firm 1 has to produce at least 3 units of output to deter Firm 2’s entry.
Therefore, even when Firm 1 acts as a monopolist and produces 4.5 units of output,
Firm 2 will not enter as 4.5 > 3. This would imply firm 1 will produce the monopolist
output of 4.5.
Hence, due to high fixed cost F, even if Firm 1 act as a monopolist and produces
monopoly output at 4.5 units of output, Firm 2 is unable to enter.