Reading
Osborne: Chapter 3
5 Auctions
A strategic Game
a set of players
for each player, a set of actions
potentially different across players
action profile: a list of actions, exactly one action from each
player.
for each player, a payoff function defined over the set of all
Nash Equilibrium
equilibrium.
An action profile in which each players action is her best
profitably deviate.
Dominance
Nash equilibrium.
When looking for Nash equilibria, we can simply eliminate
5 Auctions
Ci (qi ) = c qi .
The output by the two firms are sold at a single price,
P (Q) =
Q
0
if Q
if Q >
q2 is thus:
1 (q1 , q2 ) = q1 (P (q1 + q2 ) c)
Similarly, the profit of firm 2 is:
2 (q1 , q2 ) = q2 (P (q1 + q2 ) c)
Best Responses
max q1 (P (q1 + q2 ) c)
q1
q1 =
1
2
( c q2 ).
Best Responses
Summarizing, we have:
(
B1 (q2 ) =
1
2
( c q2 )
if q2 c
if q2 > c
B2 (q1 ) =
1
2
( c q1 )
if q1 c
if q1 > c
Nash Equilibrium
q1 =
q2 =
1
2
1
2
( c q2 )
( c q1 )
where q1 = q2 =
1
3
( c).
1
( c 0)
2
1
( c 0)
2
1
( c 0)
2
2
3
1
4
( c) , 41 ( c) .
n firms
Now suppose there are n firms, each with the same cost
n firms
function.
For example, in the case of firm 1, we solve:
max q1 (P (q1 + q2 + . . . + qn ) c)
q1
B1 (q1 ) =
1
2
( c
Pn
i=2 qi )
if ni=2 qi c
P
if ni=2 qi > c
P
Nash Equilibrium
that satisfies:
= 1 c P
q
q
i6=1 i
1
2
q = 1 c P
i6=1 qi
2
2
P
1
qn =
i6=n
qi
q1 = q2 = . . . = qn = ( c) / (n + 1).
Increasing n
n
The total output of the n firms is n+1
( c), and the
market price is
n
n+1
( c).
decreases.
As n approaches , the total output approaches ( c)
5 Auctions
D (p) =
p
0
if p
.
if p >
(pi c) ( pi )
i (p1 , p2 ) =
1
2
where j is i s competitor.
(pi c) ( pi )
if pi < pj
if pi = pj
if pi > pj
Ai = R+ ).
Payoff: each players payoff is its profit given its own price
p1 = p2 = c.
Proof:
First we need to show that (c, c) is indeed a Nash equilibrium.
Given that its opponent chooses price c, a firm can receive a
profit equal to 0.
Deviating to any price higher than c will yield a firm zero
Proof (continued):
There cannot be any Nash equilibria where at least one firm
modification?
Cournot or Bertrand?
5 Auctions
Electoral Competition
elections?
What determines the candidates policy platforms?
Why are there often no substantive differences between the
The Model
policy position, xA , xB R.
There are a continuum of voters, each has an ideal policy
position, xi R.
The ideal positions of the voters are continuously distributed
The Model
given by |y xi |.
Thus, a voter always votes for the party whose policy
The Model
They care only about winning the election (i.e., they are
office-seeking). So their payoff is assumed to be the
probability of winning.
The election is plurality rule: the party with the most votes
win.
Thus, a partys payoff is 1 (0) if strictly more (less) voters
Proof (continued):
Now consider any strategy profile where exactly one party
chooses xm .
Then that party will win the election with certainty. The
other party can increase their payoff from 0 to 21 by deviating
to xm .
Proof (continued):
Now consider any strategy profile where exactly one party
chooses xm .
Then that party will win the election with certainty. The
other party can increase their payoff from 0 to 21 by deviating
to xm .
Proof (continued):
Now suppose neither party chooses xm . Then either their
1
2
Proof (continued):
Now suppose neither party chooses xm . Then either their
1
2
Proof (continued):
Now suppose neither party chooses xm . Then either their
1
2
Proof (continued):
Now suppose neither party chooses xm . Then either their
1
2
Discussion
5 Auctions
Auctions
The Game
real number).
The payoffs to player i is vi max{bj }j6=i if her bid is among
good is given to the player among these bidders who has the
highest valuation.
Nash Equilibria
Nash Equilibria
Why?
Here player 1 also gets the good, but pays 0.
Nash Equilibria
Nash Equilibria
Nash Equilibria
Nash Equilibria
Nash Equilibria
dominated actions.
recall that strictly dominated actions can never be a part of
Nash Equilibria
Nash Equilibria
bid.
The winner is still the one with the highest bid.
However, the winner now has to pay her own bid.
When there are ties, the good goes to the player amongst the
highest bidders who has the highest valuation for the good.
Nash Equilibria
Nash Equilibria
Nash Equilibria
Nash Equilibria
Nash Equilibria
Nash Equilibria
Nash Equilibria
were the case, player 2 can bid slightly higher than the
highest bids and receive a positive payoff.
Nash Equilibria
were the case, player 2 can bid slightly higher than the
highest bids and receive a positive payoff.
Nash Equilibria
were the case, player 2 can bid slightly higher than the
highest bids and receive a positive payoff.