Midterm-Exam-November 2018-Sol
Midterm-Exam-November 2018-Sol
NAME:
X Y Z
A 0, 1 20, 1 15, 2
Problem 1: Consider the following normal form game: B 1, 0 30, 0 10 , 1
C 10, 2 2 , -1 0, 1
D 5, 1 5, 4 -1 , 5
(a) (5 points) What are the strategies that survive the iterated elimination of strictly dominated strategies?
Solution: Strategy Y is strictly dominated by strategy Z for player 2. We obtain
X Z
A 0, 1 15, 2
B 1, 0 10 , 1
C 10, 2 0, 1
D 5, 1 -1 , 5
Now, strategy D is strictly dominated by strategy C for player 1. We obtain
X Z
A 0, 1 15, 2
B 1, 0 10 , 1
C 10, 2 0, 1
The set of rationalizable strategies is {A, B, C} × {X, Z}.
(b) (5 points) Find all pure strategy Nash equilibria and the payoffs of these equilibria.
Solution: The best replies of the players are
X Z
A 0, 1 15, 2
B 1, 0 10 , 1
C 10, 2 0, 1
We obtain the NE (A, Z) with payoffs u1 = 15, u2 = 2 and (C, X) with payoffs u1 = 10, u2 = 2.
(c) (15 points) Compute the mixed strategy Nash equilibria and the expected payoffs of these equilibria.
Solution: Let us look for a NE of the form
σ1 = xA + yB + (1 − x − y)C
σ2 = pX + (1 − p)Z
The expected utilities of player 1 are
Graphically,
1
u1 (σ1 , A) = 15 − 15p
u1 (σ1 , C) = 10p
u1 (σ1 , B) = 10 − 9p
3 p
5 1
Thus,
3
x = 1, y = 0
if p< 5
3
BR1 (σ2 ) = x ∈ [0, 1], y = 0 if p= 5
3
x = 0, y = 0 if p>
5
And we see that player 1 will not use strategy B. Hence we may assume y = 0. The expected utilities of player 2
are
u2 (σ1 , X) = 1 × x + 0 × y + 2(1 − x − y) = 2 − x − 2y
u2 (σ1 , Y ) = 2x + 1 × y + 1 × (1 − x − y) = 1 + x
u2 (σ1 , X) = 2−x
u2 (σ1 , Y ) = 1+x
Hence,
1
p = 1
if x< 2
1
BR2 (σ1 ) = [0, 1] if x= 2
1
p=0 if x>
2
Graphically,
2
3
BR2 (x)
1
3
5
BR1 (p)
1
1 x
2
Problem 2: Two agents, 1 and 2, produce heterogeneous products. If the two agents set, respectively, prices p1
and p2 the quantities demanded of the products of the two agents will be,
n p2 o
x1 (p1 , p2 ) = max 0, 90 − p1 +
2
n p1 o
x2 (p1 , p2 ) = max 0, 90 − p2 +
2
3
Assume that the agents have constant marginal costs MC1 = MC2 = 30. Assume also that the characteristics of the
market are such that the agents have to set simultaneously the prices.
(a) (5 points) Write down the agents’ profits, π1 (p1 , p2 ) and π2 (p1 , p2 ), as functions of the prices they set.
Solution: Assuming that the demand is positive, the profits of the agents are
p2 p1 p2
π1 (p1 , p2 ) = 90 − p1 + (p1 − 30) = 120p1 − p21 − 15p2 + − 2700
2 2
p1 p1 p2
π2 (p1 , p2 ) = 90 − p2 + (p2 − 30) = 120p2 − p22 − 15p1 + − 2700
2 2
(b) (10 points) Find the best response of both agents: BR1 (p2 ) and BR2 (p1 ). Draw the best response function and
the Nash equilibrium in the p1 − p2 plane (p1 on the horizontal axis and p2 on the vertical axis).
Solution: The first order condition for agent 1 is
∂π1 1
= 120 − 2p1 + p2 = 0
∂p1 2
The solution is the best reply of player 1
1
BR1 (p2 ) = (p2 + 240)
4
Note that the second order condition is fulfilled.
∂ 2 π1
= −2 < 0
∂p21
Similarly, best reply of player 2 is
1
BR2 (p1 ) = (p1 + 240)
4
Graphically,
p2 BR1 (p2 )
BR2 (p1 )
80
60
60 80 p1
(c) (10 points) Find the Nash equilibrium (p∗1 , p∗2 ) and the equilibrium profits π1 (p1 ,∗ p∗2 ) and π2 (p∗1 , p∗2 ).
Solution: The NE is the solution to the equations
That is
1
p1 = (p2 + 240)
4
1
p2 = (p1 + 240)
4
4
The NE is
p∗1 = p∗2 = 80
The profits are
π1 (p∗1 , p∗2 ) = π2 (p∗1 , p∗2 ) = 2500
1.1
A B
(1, 4) 2.1
X Y
(4, 1) (0, 0)
(a) (5 points) Write the set strategies for each player. Write the normal form of the game.
Solution: The set of strategies for the players are S1 = {A, B}, S2 = {X, Y }. The game may be represented by
the table
Player 2
X Y
Player 1 A 1, 4 1, 4
B 4, 1 0, 0
σ1 = pA + (1 − p)B
σ2 = qX + (1 − q)Y
u1 (A, σ2 ) = 1
u1 (B, σ2 ) = 4q
u2 (σ1 , X) = 1 + 3q
u2 (σ1 , Y ) = 4q
5
q
BR2 (p)
1
BR1 (q)
4
1 p
We see that there are two NE, (A, Y ), with payoffs u1 = 1, u2 = 4 and (B, X), with payoffs u1 = 4, u2 = 1, in
pure strategies and the following NE
1
(A, qX + (1 − q)Y ) 0≤q≤
4
with payoffs u1 = 1, u2 = 4, in mixed strategies.
(c) (5 points) Find the subgame perfect Nash equilibrium of the game.
Solution: There are two subgames. The NE of the subgame that starts in node 1.2 is X. Thus, the SPNE is
(B, X) with payoffs u1 = 4, u2 = 1.
Problem 4: Suppose that two agents are negotiating an agreement. If the two parties reach an agreement immedi-
ately, each will have a gain of 200. If the two parties don’t achieve an agreement immediately they will have a second
and last chance of achieving an agreement in a month. If the agreement is reached in a month, agent A will have a
gain of 80 and agent B will have a gain of 90. Negotiation works in the following way: In the first period agent A
demands a payment of agent B in order to sign the agreement. If agent B accepts, the agreement is reached, the
payoff to agent A will be 200 plus the payment that agent B has accepted to make. And the payoff to agent B will
be 200 minus the payment it has accepted to make. If agent B rejects the offer, it will make an offer to agent A in
the second period with the offer consisting again of a demand of a payment in order to sign the agreement.
(a) (15 points) Describe the situation as an extensive form game. How many subgames does it have?
Solution: The extensive game form representation is the following.
6
A.1
d1
B.1
A R
d2
A.2
A R
(uA = 80 − d2 , uB = 90 + d2 ) (0, 0)
(b) (10 points) Conpute the subgame perfect Nash equilibrium of this game. Write the strategies of the players at
each node. Write the payoffs of the players in the SPNE.
Solution: At node A.2, firm A accepts iff 80 − d2 ≥ 0. That is, firm A accepts iff d2 ≤ 80. Given the best reply
of firm A at node A.2, the best response of firm B at node B.2 is to offer d2 = 80. Thus, we may assume that
if we ever reach node B.2, firm B will offer d2 = 80 and firm A accepts. The payoffs will be uA = 0, uB = 170.
We replace this payoffs at node B.2
7
A.1
d1
B.1
A D
Now, player B at node B.1 accepts iff 200 − d1 ≥ 170. That is, at node B.1 player B accepts iff d1 ≤ 30. The
best response now for player A is to offer d1 = 30 at node A.1 and player B accepts. The SPNE is the following.
• Node A.1: d1 = 30.
• Node B.1: accept iff d1 ≤ 30.
• Node B.2: d2 = 80.
• Node A.2: accept iff d2 ≤ 80.