LAST NAME: NAME:

UNIVERSITY CARLOS III

Master in Economics Master in Industrial Economics and Markets
Game Theory
MIDTERM EXAM–October 28th, 2022

Instructions: Total time. 2 hours. This is NOT an open book exam. Calculators are allowed. Write your work on
this booklet. Unsupported unswers will receive little credit.

Problem 1: Agents 1 (kettledrums) and 2 (trumpet) like to play music together. For this, they have to practice
separately, so that when they meet each one knows well his part. If agent 1 spends x1 hours practicing his part and
agent 2 spends x2 hours practicing her part, their utilities are
√
u1 (x1 , x2 ) = x1 x2 − x21
√
u2 (x1 , x2 ) = x1 x2 − x22

(a) (10 points) What is the most favorable agreement that maximizes the welfare of both agents? What are the
utilities of each agent in this agreement? (You may assume the solution is symmetric x1 = x2 )
Solution: The players maximize the sum of the utilities
√
max u1 + u2 = 2 x1 x2 − x21 − x22
x1 ,x2

The first order conditions are

x2
0 = √ − 2x1
x1 x2
x1
0 = √ − 2x2
x1 x2

We obtain solution x1 = x2 = 12 . The utilities of the players are

   
1 1 1 1 1
u1 , = u2 , =
4 4 4 4 4

(b) (10 points) Suppose from now on, that after the meeting both agents decide the time they practice x1 and x2
simultaneously and without knowing what the other one is doing. Find the best response function of each agent,
BR1 (x2 ) and BR2 (x1 ).
√
Solution: Player 1 maximizes u1 (x1 , x2 ) = x1 x2 − x21 The first order condition is
x2
√ − 2x1 = 0
2 x1 x2

The best reply of player 1 is

 x  13
2
x1 (x2 ) =
16
Similarly, the best reply of player 2 is
 x  13
1
x2 (x1 ) =
16

1
(c) (5 points) Why the agreement reached in part (a) will not be implemented if each agent takes its decision
simultaneous and independently?
Solution: Because     13
1 1 1
BR1 = ̸=
2 32 2
    13
1 1 1
BR2 = ̸=
2 32 2

(d) (10 points) Find the Nash equilibrium (x∗1 , x∗2 ) if the agents take their decisions simultaneous and independently.
Compute the utilities of the agents in the NE. (You may assume the solution is symmetric x1 = x2 )
Solution: The NE is the solution of
 x  13  x  31
2 1
x1 = , x2 =
16 16
That is
1
x∗1 = x∗2 =
4
The utilities of the players are    
1 1 1 1 3
u1 , = u2 , =
2 2 2 2 16

Problem 2: Consider the following dynamic form game.

1.1

1.2 E F 2.1

A B U V1.3
2.2 2
10, 3
X Y X Y C D
2.3
1, 0 16, 111, 1 1, 0
R S R S
0, 3 1, 21, −12, 0

(a) (20 points) Write the normal form of the game that starts at node 1.2 and find all (that is, in pure and mixed
strategies) NE of this game.
Solution: The normal form game is

X Y
A 1,0 16 , 1
B 11 , 1 1,0

In problem 1 we have computed the NE of this game.

(a) (A, Y ) with payoffs u1 = 16, u2 = 1 ;

(b) (B, X) with payoffs u1 = 11, u2 = 1; and
(c) 12 A + 12 B, 53 X + 35 Y , with payoffs u1 = 7, u2 = 12 .

(b) (10 points) Write the normal form of the game that starts at node 1.3 and find all (that is, in pure and mixed
strategies) NE of this game.
Solution: The normal form game is

2
 R S
C 0,3 1,2
D 1 , -1 2,0

Strategy C is (strictly) dominated by D,

R S
D 1 , -1 2,0

Now, strategy S is dominated by S,

S
D 2,0

There is a unique equilibrium with payoffs u1 = 2, u2 = 0.

(c) (20 points) Find the subgame perfect Nash equilibria of the complete game. Compute the utilities attained by
the players in each SPNE.
Solution: At node 2.1 , player 2 chooses U . The payoffs at the sub-game starting at node 2.1 are u1 = 10,
u2 = 23 .

1.1

E F
1.2 2
10, 3

A B

X Y X Y

1, 0 16, 1 11, 1 1, 0

(a) Let us look for a SPNE in which at node 1.2 the NE (A, Y ) is played.

1.1

E F

2
16, 1 10, 3

At node 1.1 player 1 chooses E. We have the SPNE: (E, (A, Y ), U, (D, S)), with payoffs u1 = 16, u2 = 1.
(b) Let us look for a SPNE in which at node 1.2 the NE (B, X) is played.

1.1

E F

2
11, 1 10, 3

At node 1.2 player 1 chooses E. We have the SPNE (E, (B, X), U, (D, S)), with payoffs u1 = 11, u2 = 1.
(c) Let us look for a SPNE in which at node 1.2 the NE 21 A + 12 B, 35 X + 53 Y is played.

3
 1.1

E F

1 2
7, 2
10, 3

1
+ 12 B, 35 X + 35 Y , U, (D, S) , with payoffs



At node 1.2 player 1 chooses F . We have the SPNE F, 2A
u1 = 10, u2 = 23 .

Problem 3: Two agents engage in a sequential bargaining process as follows:

1. In stage 1, agent A proposes 0 ≤ x ≤ 300.

(a) If agent B accepts, agent A gets x, agent B gets 300 − x.

(b) If agent B refuses, they go into the stage 2.
2. In stage 2, agent B proposes 0 ≤ y ≤ 200.
(a) If agent A accepts, agent A gets 200 − y, agent B gets y.
(b) If agent A refuses, agent A gets the amount c and agent B gets the amount d.
Assume c + d < 100.
(a) (10 points) Describe the situation as an extensive form game. Draw the tree that represents the game.
Solution:

A.1

B.1

A R
B.2

(uA = x, uB = 300 − x)

A.2

A R

(uA = 200 − y, uB = y) (c, d)

There are infinitely many subgames.

4
(b) (10 points) Compute the subgame perfect Nash equilibria of this game. Write the strategies of the players at
each node. Write the payoffs of the players in each of the SPNE.
Solution: At node A.2, agent A accepts iff 200 − y ≥ c. That is agent A accepts iff y ≤ 200 − c. Given the best
reply of agent A at node A.2, the best response of agent B at node B.2 is to offer y = 200 − c. Thus, we may
assume that if we ever reach node B.2, agent B will offer y = 200 − c and agent A accepts. The payoffs will be
uA = c, uB = 200 − c > d. We replace this payoffs at node A.2

A.1

B.1

A R

(uA = x, uB = 300 − x) (uA = c, uB = 200 − c)

Now, player B at node B.1 accepts iff 300 − x ≥ 200 − c. That, is at node B.1 player B accepts iff x ≤ 100 + c.
The best response now for player A is to offer x = 100 + c at node A.1. The SPNE is the following.

• Node A.1: x = 100 + c.

• Node B.1: accept iff x =≤ 100 + c.
• Node B.2: y = 200 − c.
• Node A.2: accept iff y ≤ 200 − c.

The payoffs are uA = 100 + c, uB = 200 − c.

(c) (5 points) Is there a player who wins/looses if c increases? Which one(s)?
Solution: Since, the payoffs are uA = 100 + c, uB = 200 − c, agent A is better off and agent B is worse off.
(d) (5 points) Is there a player who wins/looses if d increases? Which one(s)?
Solution: Since, the payoffs are uA = 100 + c, uB = 200 − c, the value of d plays no role.
(e) Imagine now that, if, in the second stage agent A refuses, they go into a third stage:
3. In stage 3, agent A proposes 0 ≤ z ≤ 100. If agent B accepts, agent A gets z, agent B gets 100 − z. If agent
B refuses, agent A gets the amount c and agent B gets the amount d.

(a) (20 points) Describe the new dynamic game and compute the SPNE.
Solution: The SPNE is the following.
• Node A.1: x = 200 − d.
• Node B.1: accept iff x =≤ 200 − d.
• Node B.2: y = 100 + d.
• Node A.2: accept iff y ≤ 100 + d.
• Node A.3: z = 100 − d.
• Node B.3: accept iff z ≤ 100 − d.

5
 The payoffs are uA = 200 − d, uB = 100 + d.
A.1

B.1

A R

B.2
(uA = x, uB = 300 − x)

A.2

A R

A.3
(uA = 200 − y, uB = y)

B.3

A R

(uA = z, uB = 100 − z) (uA = c, uB = d)

(b) (5 points) Is there a player who wins/looses if c increases? Which one(s)?

Solution: Since, the payoffs are uA = 200 − d, uB = 100 + d, the value of c plays no role.
(c) (5 points) Is there a player who wins/looses if d increases? Which one(s)?
Solution: Since, the payoffs are uA = 200 − d, uB = 100 + d, agent B is better off and agent A is worse
off.
(d) (10 points) Depending on the parameters c, d who prefers the 2-stage bargaining procedure versus the
3-stage bargaining procedure?
Solution:
Since, c + d < 100, we have that 100 + c < 200 − d and 200 − c > 100 + d. Therefore, A prefers the two-stage
game and agent B prefers the three-stage game.

Problem 4: Consider a market with two firms and a homogeneous product. The inverse demand function is
p = 150 − 3(q1 + q2 ), where q1 is the output of firm 1 and q2 is the output of firm 2. The firms have constant

6
marginal cost c = 6. Firm 1 (the leader) chooses its output level q1 first. Then, firm 2 (the follower) chooses its
output level q2 , after observing q1 .
(a) (10 points) Describe the situation as an extensive form game. Write the payoffs of the firms.
Solution: The set of player is N = {1, 2}. The strategy set for player 1 is S1 = [0, ∞). The strategy set for
player 2 consists of functions q2 (q1 ) : [0, ∞) → [0, ∞). The associated tree is
F1

q1

F2

q2

(π1 , π2 )

The profits of the firms are

π1 (q1 , q2 ) = (150 − 3q1 − 3q2 ) q1 − 6q1 = 144q1 − 3q12 − 3q1 q2
π2 (q1 , q2 ) = (150 − 3q1 − 3q2 ) q2 − 6q2 = 144q2 − 3q1 q2 − 3q22

(b) (20 points) Compute the subgame perfect Nash equilibria of this game. Write the strategies of the players at
each node. Write the payoffs of the players in each of the SPNE.
Solution: Firm 2 maximizes its profit function. The first order condition is
−3q2 − 3(−48 + q1 + q2 ) = 0
Firm 2’s best response to a price set by firm 1 is
 
48 − q1
BR2 (q1 ) = max 0,
2
Now, in taking its decision, Firm 1 anticipates the reaction of Firm 2. Firm 1 maximizes
3q 2
   
48 − q1 48 − q1
π1 q1 , 2
= 144q1 − 3q1 − 3q1 = 72q1 − 1
2 2 2
The first order condition is
72 − 3q1 = 0
and we obtain q1S = 24, q2S = BR2 (24) = 12. The profits are π1S (24, 12) = 864 and π2S (24, 12) = 432.
(c) (20 points) Suppose now that firms decide simultanous and independently the quantities q1 and q2 . That is,
assume that now firm 2 does not observe q1 when it decides q2 . Compute the Cournot-Nash equilibrium of the
resulting game. As compared with the previous situation, is there firm which is better/worse off? Which one(s)?
In doing your computations, you may assume the Cournot-Nash is symmetric.
Solution:    
48 − q2 48 − q1
BR1 (q2 ) = max 0, BR2 (q1 ) = max 0,
2 2
The Nash-Cournot equilibrium is the solution of the following system of equations
48 − q2 48 − q1
q1 = , q2 =
2 2
The solution is q1∗ = q2∗ = 16. The profits of the firms are π1∗ (16, 16) = π2∗ (16, 16) = 768. Firm 1 is better off
with the Stackelberg setting. Firm 2 is better off with the Cournot setting.