Unit 10 Non-Collusive Oligopoly: 10.0 Objectives
Unit 10 Non-Collusive Oligopoly: 10.0 Objectives
Unit 10 Non-Collusive Oligopoly: 10.0 Objectives
Structure
10.0 Objectives
10.1 Introduction
10.2 Non-Collusive Oligopoly
10.2.1 Cournot Model of Duopoly
10.2.2 Bertrand Model of Duopoly
10.2.3 Edgeworth Model
10.2.4 Chamberlin’s Oligopoly Model
10.2.5 Kinked Demand Curve: Sweezy Model
10.2.6 Stackelberg Model
10.3 Let Us Sum Up
10.4 Key Words
10.5 Some Useful Books
10.6 Answer or Hints to Check Your Progress
10.7 Exercises
10.0 OBJECTIVES
After going through this unit, you will be able to:
10.1 INTRODUCTION
An oligopolistic market is characterised by the existence of a small number of
firms who have the market power in the sense that they can affect the market
price by changing their output level. In such a market, the firms may produce
identical or differentiated products. The distinguishing feature in it is
strategic interdependence among the firms with regard to price and output
decisions.
strategic variable, then the CV of the ith firm would be given by (δqj/δqi)– the
amount of change in the output level that would be brought about by the jth
firm for an additional change in the output level of the ith firm, as perceived by
the ith firm. Depending on CV, we can have different models under oligopoly.
For instance, in the Cournot Duopoly model, CV of each firm is zero because
each of the duopolists assumes that the other would stick to its previous
period’s output level. In the Stackelberg model, there is a leader and a
follower. Here the leader knows what the follower is likely to do; hence, the
CV of the leader is positive.
In the following sections, we would see how equilibrium is arrived at in the
important models of non-collusive oligopoly—Cournot model of duopoly,
Bertrand model, Stackelberg model, Edgeworth, Chamberlin and the Kinked
Demand curve analysis of Sweezy. To do this we would make use of the
concept of reaction functions (RF). A reaction function of a firm gives the best
response of the firm, given the decision taken by the rival firm.
Assumptions
1) Each of the firms faces a linear market demand curve
2) Both sell identical products. In Counot’s model, the two are assumed to
sell mineral water.
3) The cost functions are identical and the marginal cost (MC) of each firm
is zero.
4) Each firm assumes that the other would continue to produce the same
output as in the last period.
Diagrammatic Representation
To arrive at the Cournot solution, let us assume that firm A is the first to
produce and sell in the market.
Let D1 D1 be the linear market demand curve, as shown in Figure 10.1.The
marginal cost =0 for both the firms. In the figure, this corresponds to the
horizontal axis. Firm A being a profit maximiser, equates MR with MC and
arrives at the output level OA (= ½ OD1) and price OP1
Suppose now firm B enters the market. As firm A is already selling OA
amount of output, firm B would cater to OD1 minus OA amount of the market
demand, assuming that firm A will continue producing OA. Therefore, the
portion of the market demand relevant to firm B is CD1.This is so because B
cannot sell anything at a price higher than OP1, as firm A is already present in
the market and they are selling the same product. Hence the only other option
open to firm B is to sell at a price lower than OP1, whereby the market demand
curve for B shrinks to CD1. Firm B being a profit maximiser, produces output
AB (= ½ A D1) where MRB = MC = 0.
6
The output level supplied in the market after firm B’s entrance is OA + AB Non-Collusive Oligopoly
(= ½ OD1 + ¼ OD1) = ¾ OD1. As the output level goes up, the price in the
market goes down to, say OP2.
Firm B now is assuming that firm A will continue producing OA1 and has ED1
as the relevant market demand curve. Setting MRB = MC = 0 firm B would
produce A1B1 = ½ A1 D1 = ½ *(1- 3/8 ) OD1 = 5/16 OD1.
The total supply in the market would be OA1 + A1B1= 3/8 OD1 + 5/16 OD1 =
11/16 OD1. As the market supply goes up the price comes down to say, OP3.
Thus, we see that the output of firm A goes down whereas that of firm B goes
up. This process continues until each one of the firms produces 1/3 OD1.To
see how, we derive the equilibrium output levels of each firm in the following.
Let the total market demand be x units of output.
Output levels of firm A:
Period 1: x/2
Period 2: ½(1-1/4) x = 3x/8 = x/2 – x/8
Period 3: ½(1- 5/16) x = 11x/32 = x/2 – x/8 – x/32
Period 4: ½(1 – 21/64) x = 43x/128 = x/2 – x/8 – x/32 –x/128
In the nth period, the output of firm A is
= x/2 – x/8 – x/32 –x/128 - …….
= x/2 – [1/8 + 1/8 * (¼) + 1/8*(1/4)2 + ……… ] * x
= x/2 –1/8 * [ 1/(1- 1/4)]*x
= x/3
Output levels of firm B:
Period 2: ½(1/2)x = x/4
Period 3: ½(1- 3/8)x = 5x/16 = x/4 + x/16
Period 4: ½(1 – 11/32) x = 21x/64 = x/4 + x/16 + x/64
Period 5: ½(1 – 43/128) x = 85x/256 = x/4 + x/16 + x/64+x/256
In the nth period, the output of firm B is
= x/4 + x/16 + x/64+x/256 +...
= [(¼)/1- ¼] * x
= x/3
7
Price and Output
Determination-II
D1
D2
MR
2
C
A
P1
MR
2
1
C
P2
MR
1
B
O A1 A B1 B D1
Fig. 10. 1: Demand Analysis of Cournot Equilibrium
In his duopoly model, Cournot makes a very naïve assumption that the firms
think their rivals would stick to their past periods output level. Therefore, the
conjectural variation (CV) of both the duopolists is equal to zero. Retaining
the same assumptions that both the duopolists i) face linear market demand
curve, ii) maximise profit and iii) have MC = 0.We can write the model as
follows:
Let the demand function be p = a - bq, where q = (qi + qj) = total market
demand and a, b > 0
Given the above assumptions, we can write the profit function of the ith firm
as:
Пi = pqi – C (qi); where i = A, B
= (a – bq) qi – C(qi)
= [a – b(qi + qj) ] qi – C(qi)
Each firm being a profit maximiser, we would differentiate Пi partially with
respect to qi and set the derivatives equal to zero.
Thus, δПi/ δqi = a – 2bqi –b (qj + qi δqj/δqi) – δC/ δqi = 0.
As in this model CV = 0, δqj/δqi = 0. Hence, we have,
a – 2bqi –bqj = 0 (as δC/ δqi = 0, by assumption)
From such an optimisation exercise we get:
qi* = (a – bqj) / 2b, qi* = Ri (qj) where, qi* gives the profit maximising level
8 of output of firm i(i, j =A, B; i ≠j)
The equilibrium output levels of both the firms is obtained by solving the two Non-Collusive Oligopoly
reaction function equations as:
qA* = a/3b, qB* = a/3b
For each firm, Ri represents the reaction function. Given the output level of
the jth firm the reaction function shows the best response (i.e., qi*) of the ith
firm, which maximises its profit. The reaction functions in this exposition will
be downward sloping straight lines, as shown below in Figure 10.2 where SP
is the reaction function of firm B and MN is the reaction function of firm A.
For any level of output of firm B say qB1 the level of output which would
maximise firm A’s profit is given by qA1 from A’s reaction function MN .We
can similarly explain each point on the reaction curve of firm B.
qB
M
Km E
qB
q1B
O q*A qA1 N P qA
Fig. 10.2: Cournot Equilibrium through Reaction Function
Equilibrium
Diagrammatically this is shown in Figure 10.2. The equilibrium is obtained
where the two reaction functions intersect each other i.e., at point E
corresponding to which we have the two equilibrium output levels as qA*
and qB*
Stability
For stability analysis, let us consider Figure 10. 3 where MN and SP represent
the reaction functions of firms A and B respectively. To see if point E is a
stable equilibrium, we would start from an arbitrary point and see if the inbuilt
dynamics of the model would take us to point E. In the diagram, ON is the
monopoly output of firm A and OS represents that of firm B. Suppose firm A
enters the market first and produces the monopoly output ON. Next, firm B
enters and assuming that firm A would continue producing at ON chooses to
produce Ob from it’s reaction function SP. Therefore, the two firms end up at
point 1 on B’s reaction function. As at point 1 firm A is off it’s reaction
function, it would not be maximising it’s profit. Hence, A would not choose to
be at 1. Assuming that firm B would continue producing at Ob, firm A would
choose to be at point 2 on it’s reaction function whereby it would be
producing Oa amount of output. From the point of view of firm B, point 2 is
9
Price and Output not a desirable one because given that A produces Oa, Ob is no more the
Determination-II
profit maximising level of output. Hence, B would move from point 2 to point
3 on it’s own reaction function. Finally, such actions would take both the
firms to point E. This movement towards E would occur if we start from a
point like OS above the equilibrium point. Thus, we see that point E is a stable
equilibrium point because any disturbance from point E would take us back
to E.
qB
M
E
3
2 1
b
O a N P qA
Fig. 10. 3: Stability Analysis in Cournot Model
……………………………………………………………………………..
……………………………………………………………………………..
……………………………………………………………………………..
……………………………………………………………………………..
……………………………………………………………………………..
…………………………………………………………………………
…..
……………………………………………………………………………..
……………………………………………………………………………..
……………………………………………………………………………..
……………………………………………………………………………..
10
3) What is a reaction function? Non-Collusive Oligopoly
……………………………………………………………………………..
……………………………………………………………………………..
……………………………………………………………………………..
……………………………………………………………………………..
……………………………………………………………………………..
……………………………………………………………………………..
……………………………………………………………………………..
……………………………………………………………………………..
……………………………………………………………………………..
……………………………………………………………………………..
Diagrammatic Representation
In the following figures, we represent the process by which the firms reach
equilibrium in the Bertrand model. The horizontal axis measures quantity of
output produced by each firm along with cost (assumed to be identical). Along
the vertical axis, we measure the price and marginal cost. In each of the
figures, the horizontal line C represents the marginal cost (MC) of each firm.
When PA is greater than PB and both are greater than C as in Figure 10. 4, firm
A will be loosing the market to firm B as they are selling homogeneous
product. Therefore, profit of firm A will be less than that of B and firm A will
undercut firm B. In the next period firm B will be loosing the market to firm
A and therefore, would undercut firm A. This process will go on until both 11
Price and Output charge the same price, as shown in Figure 10. 5. In such a case, both of get an
Determination-II
equal share of the market. However, it will not be an equilibrium situation
because if any one of them reduces its price marginally, then it gets the full
share of the market and earns more profit than when they were sharing the
market equally. Finally, equilibrium will be attained when PA = PB = MC.
This is shown in Figure 10. 6.
Figure 10. 6 gives a stable equilibrium because no one has any incentive to
reduce or raise the price. If the former occurs, then there is the threat of loss
because MC is greater than price. If the price is higher than the equilibrium
price (P*), then there is a threat of potential loss of customers. Therefore, if
price is greater than or less than P* it will have a tendency to move back to
P*, which explains the stability of the equilibrium.
PA
PA =PB
PB
C C P* =C
1) Describe how the firms under Bertrand model arrive at the equilibrium
price.
………………………………………………………………………………
………………………………………………………………………………
………………………………………………………………………………
………………………………………………………………………………
………………………………………………………………………………
12
Non-Collusive Oligopoly
Unlike Bertrand, Edgeworth assumes that none of the duopolists can produce
an output as large as the competitive market’s.
Suppose firm A is the first one to enter the market. Equating MR and MC, it
decides to produce OE at price OP1.Now suppose B enters the market and sets
price slightly below OP1 and thus captures all of A’s customers. But then B
can cater to the market demand only up to OD, hence amount left to be sold
by firm A is AC (= DB, by construction). Firm A rather than accept the
reduction in revenue decides to reduce the price slightly below that of B’s. As
a result, A can now capture all of B’s customers. However, once again, A can
sell only up to OC. This process would go on until price falls to OP2, and each
of the duopolist produce the maximum possible output.
However, the price OP2 is not stable because one of the firms can raise it’s
price and thereby it’s revenue as well as profit (because MC = 0). For
instance, A will try to raise price assuming B will maintain it if OP2. A has no
fears of loosing customers to B, because B is producing it’s maximum
possible output so that rest of the market is to be catered to by A only. Now B
would also reconsider that a price rise from OP2 would not result in a loss of
sales. Therefore, it would raise price almost up to OP1. A would respond by
reducing price and the same process ensues once again as before. We see that
price would fluctuate between OP1 and OP2 and there would be no stable price
equilibrium.
………………………………………………………………………………
………………………………………………………………………………
13
Price and Output 2) Is there a stable equilibrium price in Edgeworth model? Why?
Determination-II
………………………………………………………………………………
………………………………………………………………………………
………………………………………………………………………………
………………………………………………………………………………
………………………………………………………………………………
1) Describe briefly the method by which the firms arrive at the equilibrium
price and output in Chamberlin’s model.
………………………………………………………………………………
………………………………………………………………………………
………………………………………………………………………………
………………………………………………………………………………
………………………………………………………………………………
………………………………………………………………………………
………………………………………………………………………………
………………………………………………………………………………
………………………………………………………………………………
………………………………………………………………………………
………………………………………………………………………………
………………………………………………………………………………
Each firm in an oligopolistic market faces two demand curves D1D1 and D2D2
as shown in Figure 10.9. D1D1 is the demand curve that a particular oligopolist
faces on the assumption that others do not change their prices and D2D2 has
been drawn on the assumption that if one firm changes the price, then all
others also change theirs.
15
Price and Output
Determination-II
See that there is some range within which changes in the firm’s marginal cost
will not result in changes in price and quantity. This is shown in Figure 10. 11.
Note that both for MC1 and MC2, the price and quantity given by the
equilibrium condition MC = MR are the same. Hence, the price is “sticky”,
and the model is also known as the “sticky price model”.
Fig. 10. 11: MR and MC Curves with Sticky Price Kinked Demand Model
The kinked demand is derived on the assumption that price increase by one of
the oligopolistic firm is not followed by others but price reductions.
………………………………………………………………………………
………………………………………………………………………………
………………………………………………………………………………
………………………………………………………………………………
………………………………………………………………………………
………………………………………………………………………………
………………………………………………………………………………
17
Price and Output 2) Describe the two different demand curves faced by a firm. Derive the MR
Determination-II
curve diagrammatically.
………………………………………………………………………………
………………………………………………………………………………
………………………………………………………………………………
………………………………………………………………………………
………………………………………………………………………………
………………………………………………………………………………
………………………………………………………………………………
………………………………………………………………………………
………………………………………………………………………………
………………………………………………………………………………
• Demand function faced by both the firms are p = p (q), where q = qA+qB ,
the aggregate output produced by both the firms.
• Cost function Ci = Ci (qi) where i = A,B
Derivation of Iso-profit Curve
18
As П0 is a constant, dП0 = 0 and pdqA + qA dp/dq (dqA + dqB) – ∂C/∂qA dqA = Non-Collusive Oligopoly
0. By simple manipulation, we see that
dqB/dqA = - [(MRA + MCA)/(qA dp/dq)]. The demand curve by assumption is
downward sloping, therefore dp/dq < 0. Hence, the iso-profit curve will be > =
< 0 according as MRA> = < MCA.
qA remaining constant at q1Aif qB increases from q1B to q2B , the firm ends up at
the iso-profit curve П0, which represents a lower level of iso-profit than П1.
Hence, the iso-profit curves lying away from the horizontal axis represents
lower level of profit.
The Stackelberg model of oligopoly deals with the leadership of a firm. Let
firm A be the leader which implies that it will make a conjecture that firm B
will accept A’s output as a datum while making a output plan and B will
actually behave in this manner. In other words, firm A will incorporate this in
the profit maximising objective of firm B. For this reason, which the
conjectural variation of firm B is zero, it is non-zero for firm A. Basically,
firm B operates on the naïve Cournot conjecture. In other words, firm A being
the leader can perceive the reaction function of firm B and would know before
hand the strategy firm B is going to adopt. Firm A incorporates this
knowledge in its profit maximising exercise in the sense that it knows the
19
Price and Output reaction function of B. Given the iso-profit curves of A and the reaction
Determination-II
function of B, firm A can find out from the tangency between the two. This is
illustrated in the following diagram. In the diagram, point E gives the
Stackelberg equilibrium. Firm B being on its reaction function would have no
incentive to deviate from E. In addition, as firm A is maximising its profit, it
has no incentive to deviate. Hence, this equilibrium is stable.
Numerical Example
Suppose the demand and cost conditions are the same as in above with firm 1
the leader and firm 2 the follower. In that case, derive the Stackelberg
equilibrium quantity and profits of each firm.
= 47.5q1 – 0.25q 12
………………………………………………………………………………
………………………………………………………………………………
………………………………………………………………………………
………………………………………………………………………………
………………………………………………………………………………
………………………………………………………………………………
………………………………………………………………………………
………………………………………………………………………………
………………………………………………………………………………
………………………………………………………………………………
The Cournot model deals with the case when the firms make conjecture that
the rival would stick to the previous level of output. Here the firms deal with
output changes. Finally, the firms together end up producing 2/3 of the total
market demand.
In the Bertrand model, the case is similar to that of Cournot except that the
firms compete in terms of price. Here they end up producing the competitive
level of output.
In the Stackelberg model one firm acts as the leader and the other follower. A
firm is a leader in the sense that it knows the reaction function of the follower.
The leader maximises profit after incorporating the reaction function of the
follower.
Mutual Interdependence: The action of one firm in the market affecting that
of the other.
21
Price and Output Price Competition: When firms compete in terms of price i.e., price is the
Determination-II
strategic variable.
Reaction Function: Given the output level of the jth firm the reaction function
shows the best response of the ith firm, which maximises the ith firm’s profit.
Sticky Price Model: Model where the price and also output remain
unchanged within a certain range.
10.7 EXERCISES
1) Check that if the reaction curves are reversed then the system would have
an unstable equilibrium in Counot’s duopoly model.
2) Suppose the duopolists face the following demand and cost functions
P = 100 – 0.5(q 1 + q 2) and Ci = 5 q i , where i = 1,2
Derive the Cournot equilibrium quantities and price and profit (of both the
firms).
3) Describe what happens in the Bertrand model when the identical cost
structure assumption is relaxed.
4) If the demand function is given by P = 100 – 0.5(q 1 + q 2) and the cost
functions of the two firms are C1 = 5q1 and C2 = 0.5q22 respectively then
find the equilibrium quantity and price charged by the firms if (a) firm 1 is
the leader (b) firm 2 is the leader.
5) In the Edgeworth model will an increase in demand raise price? Why?
6) In the kinked demand model of oligopoly at what price does the kink
occur? How useful is the model in explaining pricing under oligopoly?
7) Discuss the welfare aspects of each model in terms of the consumers in
the society.
[Hint: Assume that the welfare of the consumers in an economy depends
on the amount of output being produced]
23