Lecture Notes: Oligopoly: Intermediate Microeconomics
Lecture Notes: Oligopoly: Intermediate Microeconomics
Lecture Notes: Oligopoly: Intermediate Microeconomics
Intermediate Microeconomics
Spring 2012
Oligopoly markets consist of a small number of rms that sell dierentiated or homogenous
products. These are markets that are dominated by a few sellers of a product. From now
on, we will focus on the simple case which is a market with only two rms, a duopoly. But
all the results apply to the more general case of oligopolies. We want to see what kind of
market equilibria arise when the industry is characterized as an oligopoly/duopoly.
There are dierent ways of modeling oligopolies. These have to do with the way in which
we let the rms interact in the market. These approaches are:
1. Assume that rms choose quantity q and then price p adjusts so that demand equals
supply. These are the Cournot and Stackelberg models.
2. Assume that rms choose the prices and then consumers choose from which rm to
buy. This is the Bertrand Model.
The predictions of what the equilibrium will look like, dier drastically between these two
models. We will now analyze the Nash equilibrium of the above models.
1
1 Nash Equilibrium in the Cournot Model
Consider the following example. There are two rms in the industry, each with a marginal
cost of $10. The total demand in this market is given by the following inverse demand
function:
P = 100 q
1
q
2
Compute the Cournot equilibrium (p
, q
1
(q
2
) = 45
1
2
q
2
2. Step 2: Compute the best response of rm 2 to rm 1s decision. This is done in
2
exactly the same way as above and we get:
100 q
2
2q
1
= 10
q
2
(q
1
) = 45
1
2
q
1
3. Step 3: Compute the Cournot equilibrium which occurs at the point where the two
above best response functions intersect.
q
1
(q
2
) = 45
1
2
q
2
(q
1
)
q
1
= 45
1
2
45
1
2
q
1
1
= 30, q
2
= 30
2 Nash Equilibrium in the Stackelberg Model
There are two rms and they make their production decisions sequentially. Each rm has
a marginal cost of $10. The total demand in this market is given by the following inverse
demand function:
P = 100 q
1
q
2
Compute the Stackelberg equilibrium (p
, q
2
(q
1
) = 45
1
2
q
1
2. Step 2: Firm 1 will take into account rm 2s response when it is making its decision
about how much to produce. Therefore rm 1s total revenue, given rm 2s best
response will be:
TR
1
(q
1
, q
2
) = (100 q
1
q
2
(q
1
))q
1
(100 q
1
45 +
1
2
q
1
)q
1
So that rm 1s marginal revenue is going to be given as,
MR = 100 2q
1
45 +q
1
4
And rm 1 chooses the quantity that satises MR=MC, so that,
100 2q
1
45 +q
1
= 10
q
1
= 45
3. Step 3: Given rm 1s optimal decision, q
1
, we can determine q
2
using rm 2s best
response function:
q
2
(q
1
) = q
2
(45)
= 45
1
2
45
= 22.5
Notice that the Stackelberg outcome diers signicantly from the Cournot equilibrium
outcome (in terms of the equilibrium quantity of production for both rms). Specically,
we see that the Stackelberg leader ends up producing more in the Stackelberg equilibrium
and the follower produces less (compared to the Cournot outcome). Why? This is the
rst-mover advantage.
3 Nash Equilibrium in the Bertrand Duopoly Model
In the models of oligopoly above, the assumption was that rms compete through setting
their quantities of production. In the Bertrand model, rms compete by setting prices
instead. What will be the Nash equilibrium in such an industry? An important result
here is that, the Nash equilibrium in the Bertrand model has to be such that the prices are
equal to the marginal costs of the rms. More formally, the Nash equilibrium (p
1
, p
2
) has
5
to be such that p
1
= MC and p
2
= MC. Below we will see why such an equilibrium arises
in the Bertrand model. To do this, we will consider all possible prices that are dierent
from p
1
= MC and p
2
= MC, and show that they cannot be an equilibrium.
1. Set prices so that p < MC. This cannot be an equilibrium because at least one rm
will earn negative prots, and therefore the rms have an incentive to deviate.
2. Set prices so that MC = p
1
< p
2
. This is not an equilibrium, because rm 1 can
deviate by setting his prices so that MC < p
1
< p
2
and make higher prots. At this
new price p
1
, rm 1 doesnt lose any customers, but is selling at a higher price (i.e.
higher prots).
3. Set prices so that MC < p
1
< p
2
. This is not an equilibrium, because rm 1 has
an incentive to deviate here by setting p
1
< p
1
< p
2
, so that he does not lose any
customers, but sells at a higher price (hence getting more prots). In the same way,
rm 2 can also deviate by setting MC < p
2
< p
1
, so that he grabs the whole market
and sells at a price above his marginal cost.
4. Set prices so that MC < p
1
= p
2
. Firm 1 has an incentive to deviate here by setting
MC < p
1
< p
1
. At this new price, rm will be charging lower price, but will have
the whole market.
The above shows that both rms setting their prices equal to their marginal cost is the
Nash equilibrium of the Bertrand model. We establish this fact above by showing that
the rms setting their prices at any other value (given what the other is doing) cannot
constitute an equilibrium. Specically, we showed how each rm will have an incentive to
deviate from their chosen strategy, if we are at any point other than that where prices are
equal to the marginal costs.
6