Intermediate Microeconomics
Game Theory and Oligopoly
Game Theory

So far we have only studied situations that were not “strategic”.

The optimal behavior of any given individual or firm did not depend on
what other individuals or firms did.
 E.g.
 An individual buys something if its price is less than his
willingness to pay.
 A firm enters a market if there are positive economic profits to be
made at going prices.

Obviously, we might want to expand this.
 For example, what happens when firms recognize how price will be
affected by their behavior (i.e., not price takers)?
 Or when one firm or person’s optimal behavior depends on what
another firm or person does.
Game Theory

Game theory helps to model strategic behavior --- or interactions
where what is optimal for a given agent depends on what actions are
taken by another agent and vice versa.

Applications:








The study of oligopolies (industries containing only a few firms)
The study of externalities and public goods; e.g. using a common
resource such as a fishery.
The study of military strategies.
Bargaining.
How markets work.
Behavior in the courts.
Behavior of news media.
Crime.
What is a Game?

A game consists of:

a set of players

a set of strategies for each player
 i.e. actions to be performed given any observed state of the world

the payoffs to each player for every possible choice of actions by that
player and all the other players.
Simultaneous Move Games


Consider games where players must choose an action without
knowing what the other players have chosen.

Does a defendant agree to testify against his co-defendants when he
doesn’t know whether or not his co-defendants are going to do the
same?

How much should a firm bid for a given item in a silent auction?

Should I act friendly or defensively when I encounter a stranger on an
empty street late at night?
How do we model the outcomes in these types of games?
Simultaneous Move Games

“Prisoners’ Dilemma”


Payoffs to Player-1;



Consider a game with 2 players, each player has two options:
 Keep quiet (“cooperate” with each other),
 or talk to Police (“defect”)
Cooperate utility of 5 if Player-2 cooperates, utility of 2 if Player-2
defects.
Defect utility of 8 if Player-2 cooperates, utility of 3 if Player-2 defects.
Payoffs to Player-2 are analogous.
Simultaneous Move Games

One way to summarize the payoffs associated with each
action is to use a payoff matrix.
Player-2
C
Player-1
C
D
(5,5)
(2,8)
(8,2)
(3,3)
D
Payoff for Player-1 (row player) shown first, followed by
payoff for Player-2 (column player)
Simultaneous Move Games

So how should we think of how to model the outcome of such
games?
Player-2
C
Player-1
D
C
D
(5,5)
(2,8)
(8,2)
(3,3)

What is Player-1’s best action to take if Player-2 Cooperates?

What is Player-1’s best action if Player-2 Defects?

How about for Player-2?

So what do think each Player will do?
Simultaneous Move Games

Dominant Strategy - A strategy that gives higher utility than all
other strategies given any actions taken by other players.

Does a dominant strategy always exist?

Meeting time for dinner?

Wearing a costume to a Halloween party?
Simultaneous Move Games

“Arms Race” - Consider a game of the following form:
Player-2
ignore
ignore
Player-1
attack
attack
(0,0)
(-4,-1)
(-1,-4)
(-3,-3)

What is Player-1’s best action to take if Player-2 chooses ignore?

What is Player-1’s best action to take if Player-2 chooses attack?

How about for Player-2?
Nash Equilibrium

Nash Equilibrium – A set of actions such that each person’s action
is (privately) optimal given the actions of others.


Key to a Nash Equilibrium:
 No person has an incentive to deviate from his Nash equilibrium
action given everyone else behaves according to their Nash
equilibrium action.
Nash Equilibrium in “A Beautiful Mind?”
Simultaneous Move Games

Nash Equilibria of “Prisoner’s Dilemma”?
Player-2
C
Player-1
D
C
D
(5,5)
(2,8)
(8,2)
(3,3)

Both Cooperate?

One Cooperate, other Defect?

Both Defect?
Simultaneous Move Games

Nash Equilibria of Arms Race?
Player-2
I
Player-1
A
I
A
(0,0)
(-4,-1)
(-1,-4)
(-3,-3)

Both Ignore?

One Ignore, other Attack?

Both Attack?
Nash Equilibria


Three things to notice:

Playing Dominant Strategies are always a Nash Equilibrium
(e.g. Prisoner’s Dilemma).

Nash Equilibria do not have to be Pareto Efficient (e.g.
Prisoner’s Dilemma and Arms Race).

There can be multiple equilibria that often that can be Pareto
ranked (e.g. Arms Race).
Applications of these types of games?
Game Theory Application: Trade

Suppose Acme Corp. could make a deal with China Corp. to produce
widgets abroad.




If both stick with the deal (i.e. China Corp. produces quality widgets and Acme
Corp. pays China Corp. the agreed upon fee), Acme’s profits will be $200K
while China Corp’s profits will be $50K.
If Acme cheats and pays less than the agreed upon rate after delivery, Acme has
profits of $250K and China Corp. ends up losing $50K.
Alternatively, if Acme acts honestly, but China Corp. cheats and produces substandard widgets, Acme Corp.’s profits will only be $50K, but China Corp.’s
profits will be $90K.
If both act dishonestly, Acme will make only $75K while China Corp. will lose
$20K.

If Acme produces widgets domestically, its profits will only be $100K and
China Corp. will have profits of $0.

Should trade happen? Will trade happen?
Game Theory Application: Trade

What is the key problem that leads to inefficiency?

How could this problem be overcome?
Continuous Actions

What if players could choose among a continuum of actions?



The standard way to handle such situation is to use Reaction Functions.
Reaction function – a function that maps any possible action by Player
b into optimal action for player a.
A Nash Equilibrium will arise at the point where Reaction functions
intersect.
Continuous Actions: Cournot Equilibrium

To see the role played by reaction functions we can look at an
Oligopoly setting.

So far, we have examined 2 types of market structure.
1.
2.

Markets where each supplier was small enough that its decision
regarding how much to supply had no effect on price (competition)
Markets where there was only one supplier, so its decision regarding
how much to supply fully determined price (Monopoly)
What happens with “a few” suppliers who don’t collude (i.e. and
oligopoly)?
Continuous Actions: Cournot Equilibrium

Cournot Equilibrium

Consider a market with two firms, each with cost function equal to
C(q) = 4q

Suppose the market (inverse) demand function is
p(Q) = 84 – 8(Q)

where Q = q1 + q2
How do we find optimal quantity supplied by each firm?
Continuous Actions: Cournot Equilibrium

So each firm’s reaction function will
be:
5  q2 / 2 if q2  10
q1  
0 otherwise
5  q1 / 2 if q1  10
q2  
0 otherwise

So what will be Nash Equilibrium?

Will it be efficient?
q1
10
Firm 2’s
reaction fn
Firm 1’s
reaction fn
5
5
10
q2
Continuous Actions: Cournot Equilibrium

What happens as the number of firms increases?

Firm 1 :

FOC:
max p(Q)q1 – C(q1)

p(Q) 
p
(
Q
)

q1   MC (Q)  0

Q


 p(Q) q1 
p(Q) 1 
 MC (Q)

Q p(Q) 

 p(Q) Q q1 
p(Q) 1 
 MC (Q)

Q p(Q) Q 

 1 
p (Q) 1  s1   MC (Q)
  

divide both terms in
brackets by p(Q)
multiply second term in
brackets by Q/Q
(s1 = q1/Q or firm 1’s market share)
So what happens when s1 = 1 (i.e. firm 1 is a monopolist)? How about when then
number of firms becomes larger, so s1 gets smaller? What about when s1 goes to zero
(i.e. firm 1 is an extremely small part of the overall market)?
Continuous Actions: Bertrand Equilibrium

Is Cournot Equilibrium always a realistic model?

How else might competition work? What would happen in this case?
Continuous Actions: Bertrand Equilibrium

Bertrand Equilibrium



Instead of choosing quantity, firms choose price to sell at.
Consider again two firms with C(q) = 4q who face an inverse market
demand function of p(Q) = 84 – 8(Q) or equivalently a market demand
curve
Q(p) = 80.5 – p/8
What is the optimal strategy for firm 1 for any given action by firm
2?