Problem Set 1 Solutions

Question 1

Question 1, (a)

To find all the Pure Strategy Nash equilibria we have to use the underlying method. You will see that
there are two pure strategy nash Equilibria (U,L) and (D,R).

L R P MP
U (3,2) (0,0) (1.2,1) (0,0)
D (0,0) (2,3) (0,0) (1.4,1.5)
P (1,1.2) (0,0) (0,0) (0,0)
MP (0,0) (1.5,1.4) (0,0) (0,0)

Notes:
a. Saying (3,2) is a pure strategy Nash Equilibrium is wrong. NE refers to the strategies that the players
select, not the payoffs.
b. If you are asked to find the Nash Equilibria you have to look for both pure and mixed NE.

Question 1, (b)
For action P or MP to be strictly dominated you need to find an other action σ ∗ that is strictly better,
for all the actions of the opponent. The other action σ ∗ can be a pure or a mixed strategy. It is easy to
see that for the row player, U does not strictly dominate P (why?), that D does not strictly dominate P
(why?) and that MP does not strictly dominate P (why?).

Thus, as we cannot dominate P with a pure strategy, we need to look for a mixed strategy. Suppose the
row player plays a mixed strategy σ ∗ in which U is played with probability α and D with probability
(1 − α). Then, to strictly dominate P we need (3α > 1), (2(1 − α) > 0), (1.2α > 0), (1.4(1 − α) > 0).
If these four conditions are true, σ ∗ is strictly better than P for every action of the opponent, and thus
σ ∗ strictly dominates P. These four conditions yield 1 > α > 1/3. Thus, any σ ∗ = (α, 1 − α, 0, 0) with
1/3 < α < 1 strictly dominates P. Using this method you can show that MP is strictly dominated for the
row player (why?) and that P and MP are strictly dominated for the column player (why?).

1
Question 1, (c)
After we remove P and MP for both players we end up with the following 2x2 game. The row player
plays (p,1-p) and the column player plays (q,1-q).

q (1-q)
L R
p U (3,2) (0,0) = 3q+0(1-q)
(1-p) D (0,0) (2,3) = 0q+2(1-q)

=
2p+0(1-p) 0p+3(1-p)

2
We end up with two conditions, 3q = 2(1 − q) ⇔ q = 5 and 2p = 3(1 − p) ⇔ p = 53 . Thus, the mixed
 
Nash Equilibrium is σ ∗ = ( 53 , 52 ), ( 25 , 53 ) .

Question 2

Question 2, (a)

The player strategies are their bids, each has 5 strategies {£0, £1, £2, £3, £4}.

Question 2, (b)

In this game the winner pays the bid of the loser.

£0 £1 £2 £3 £4
£0 (1,1.5) (0,3) (0,3) (0,3) (0,3)
£1 (2,0) (0.5,1) (0,2) (0,2) (0,2)
£2 (2,0) (1,0) (0,0.5) (0,1) (0,1)
£3 (2,0) (1,0) (0,0) (-0.5,0) (0,0)
£4 (2,0) (1,0) (0,0) (-1,0) (-1,-0.5)

Question 2, (c)
There is no strict dominance in this game. For Bob, bidding £3 is weakly dominant. For Ann, bidding
£2 is weakly dominant. Notice that for both the weakly dominant strategy is equal to their valuation.

2
Note: Understand the difference between strict and weak dominance. They look similar, but they are
not identical. For example, weak dominance can eliminate Nash equilibria, strict dominance will never
eliminate Nash Equilibria.

Question 3

Question 3, (a)

The strategic form of the game is the following:

Shirk Work
Shirk (0,0) ( 21 , 12 -c)
Work ( 12 -c, 12 ) (1-c,1-c)

Question 3, (b)

Consider the row player, as the game is symmetric the same will apply for the column player as well.

1 1
For Shirk to dominate Work we need: 0 > 2 − c and 2 > 1 − c, that means c > 12 .

1
For Work to dominate Shirk we need: 2 − c > 0 and 1 − c > 12 , that means c < 12 .

Question 3, (c)

For c = 21 , we do not have strict dominance and thus this is the other case. The resulting strategic form
is:

Shirk Work
Shirk (0,0) ( 12 ,0)
Work (0, 12 ) ( 21 , 12 )

By the underlining method (Shirk, Shirk) is a NE. In fact, any action pure or mixed is a NE in this game
(why? Hint: follow the method of Q1.(c) to derive this).

3
Question 4

Question 4, (a)

To find the values of α that make the 1st firm profit function concave we need to look at the second
derivative.
Π1 = 50 − α · (a − x)2 + (1 − α)(x − y)
∂Π1
∂x = 2α(a − x) + (1 − α)
∂ 2 Π1
∂x2
= -2α

∂ 2 Π1
For concavity, we need ∂x2
≤ 0 ⇔ α ≥ 0. For the best response to exist, we need strict concavity. Weak
concavity is not enough, as a straight line with positive slope is weakly concave (more is always better
for the firm and thus there is no finite optimal choice).

Question 4, (b)

First we find the best response function of each firm.

∂Π1
∂x = 0 ⇔ 2α(a − x) + (1 − α) = 0 and for α = 12 , we have x = a + 1
2

∂Π2 b+x
∂x = 0 ⇔ 2(b − y) + 2(x − y) = 0 ⇔ y = 2

The NE is the intersection of the best response functions, thus we plug the best response of the first firm
 
in the best response of the second and the resulting NE is σ ∗ = a + 12 , b+a
2 + 1
4 . To find the equilibrium

profits plug the NE actions in the profit function of each firm.

Question 4, (c)

1 b+a 1 1
For x < y we need a + 2 < 2 + 4 ⇔a+ 2 < b.

4
Question 5

The question asks you to find a Nash Equilibrium. This can be either an SPE or a NE that is not a SPE
(and thus includes a non-credible threat). The game tree is the following:

0 1

A R

   
w, θ − w − δ, 0

To find the SPE simply do backwards induction.

1. The firm accepts if w ≤ θ and rejects otherwise.

2. The union maximizes its payoff given 1. and thus proposes w = θ

This game also has NE that are not SPE, can you find one ? (Hint: Consider non-credible threats
by the firm.) Note: It is important to understand how to write an SPE correctly. If you just write
 
σ ∗ = w = θ, f irm accepts it is wrong (why?) ← this is important.