ECON 2112.

PROBLEM SET 1

DUE: 5 PM, 26 FEBRUARY (FRIDAY)

Note: In some exercises below you will find the expression “pure strategy”. By that we mean the kind
of strategies that we have seen in the first week (particularly, in parts 1 and 2 of Week 1 Lecture Video).
Later we will introduce concept of mixed strategies and the terminology will become clear.

Exercise 1. Stag Hunt Game

Two individuals go out on a hunt. Each can individually choose to hunt a stag or hunt a hare. Each
hunter must choose an action without knowing the choice of the other. The only way of hunting a stag is
that both hunters choose to hunt it. An individual can always get a hare by himself, but a hare is worth
less than a stag. To be precise, each hunter will get a payoff equal to two for successfully hunting a stag.
A hunter will get a payoff equal to one for hunting a hare. Finally, a hunter will get a payoff equal to
zero if she does not hunt anything.
(i) Identify the set of pure strategies for each player.
(ii) Write down the normal form game.
(iii) Compute the set of pure strategy Nash equilibria.

Exercise 2. Strategic Subsidies

Airbus (a European firm) and Boeing (an US firm) are competitors in the market for aircrafts. The
demand for aircrafts has gone down significantly in recent times - in fact, so much so that only one
player can profitably operate in the market. If both stay in the market, each loses $5 million. If both exit
the market each gets 0, where as if one stays and one exits, the one who exits gets 0 while the one who
stays bags $100 million.
(i) Identify the set of pure strategies for each player.
(ii) Write down the normal form game and the associated payoff matrix.
(iii) Compute the set of pure strategy Nash equilibria.
Now suppose the US government announces and credibly commits to $6 million subsidy for Boeing,
contingent on Boeing not exiting. That is, Boeing will receive $6 million as long as it remains in the
market[note: this means $6 million gets added to Boeing’s profit/loss figures stated above provided it
stays]. How does this affect your answers to (ii) and (iii)?

Exercise 3. Definition
Provide a formal definition for the following concepts:
(i) n-player normal form game.
(ii) Nash equilibrium.
(iii) A player’s best response.

1
2

Exercise 4. Beyond two players

Consider the following 3-player normal form game

L R L R
T 8, 8, 8 7, 7, 7 T 0, 0, 9 1, 1, 8
B 7, 7, 7 8, 8, 8 B 1, 1, 8 0, 0, 9
W E
(i) For each player, write his set of best responses against every strategy profile of his opponents.
(ii) Find every Nash equilibrium in the game.

Exercise 5. Beyond two strategies

Provide an example of a 2-player normal form game where each player has 3 (pure) strategies such that:
(i) There is no pure strategy Nash equilibrium.
(ii) There are exactly three pure strategy Nash equilibria.
(iii) There are exactly nine pure strategy Nash equilibria.
 ECON 2112. PROBLEM SET 2

DUE: SATURDAY 5 PM, 6TH MARCH

*Please write your name and ID on the first page of your assignment

Exercise 1. Finding Nash equilibria

Find all pure and mixed strategy Nash equilibria of the following games.
(i)

L R
T 0, 0 1, 3
B 3, 1 0, 0
(ii)

L R
T 9, −9 7, −7
B 6, −6 8, −8
(iii)

L C R
T 4, 3 2, 7 0, 4
B 5, 5 5, −1 −4, −2

Exercise 2. Voting
Voters in Suburbia are choosing between two candidates, L and R. L is proposing new infras-
tructure funded with new taxes. Candidate R is in favour of preserving the status quo, i.e., no
new taxes and no new spending. Consider three voters - Alison, Beth, and Chandra - whose
preferences and payoff are as follows.
• Alison strictly prefers L to R. She gets a payoff of 2 if L wins but 0 otherwise.
• Beth strictly prefers R to L. She gets a payoff of 2 if R wins but 0 otherwise.
• Chandra is not interested in electoral politics. Nonetheless voting is compulsory in
Suburbia so he always votes. He gets 0 if the candidate he votes for wins. Otherwise,
his payoff is -1.

All three individuals - Alison, Beth, and Chandra - vote. Candidates receiving majority of the
votes wins.
(i) Write down the 3-person normal form game where Alison, Beth, and Chandra are three
players and each of them has two strategies - vote for L or vote for R. As described
above, their payoffs depend on the outcome of the election.
1
2

(ii) Find out all Nash equilibria in pure strategies.

(iii) Suppose we are interested in Nash equilibria where neither strictly nor weakly domi-
nated strategies are played. A Nash equilibrium satisfying this property is called ad-
missible. Are all Nash equilibria in (ii) admissible?

Exercise 3. Entry into new markets

Hunter and Melissa are two competing coffee growers who are exploring new markets to sell
their product. There are two potential markets - Large and Small - referred to as L and S
hereafter. Entry costs are large so each of them can enter only one of the two markets.
• If one of them enters L while the other one enters S, then one who enters L gets 36, while
the other one gets 9.
• If both enter L, each gets 36α.
• If both enter S, each gets 9α.
where 0 ≤ α ≤ 21 .

(i) The game described above has at least one Nash equilibrium in pure strategies irrespective
of the value of α - Do you agree?
(ii) The game described above has at least one Nash equilibrium in strictly mixed strategies
irrespective of the value of α - Do you agree? [Note: By strictly mixed strategies we refer
to strategies of the form (xL + (1 − x)S , yL + (1 − y)S ) where 0 < x < 1 and 0 < y < 1.]

Exercise 4. Price competition

There are 100 consumers in the suburb Little Italy all of whom buy either one medium pizza
or nothing. Each consumer has a reservation price of $15. That is, each consumer is willing
to pay at most $15 for a medium pizza. There are two pizza stores - A and B - located next to
each other. It costs $7 to make a medium pizza. The price of pizza is in whole dollars (e.g., $8,
$7, but not $6.99 or $8.50)
Let pa and pb denote the price charged by stores A and B respectively. For simplicity, assume
pi ∈ {1, 2, ..., 20} for both i = A, B. If pi < p j , all consumers buy pizza from store i as long as
pi ≤ $15. If pi = p j = p, and p ≤15, then50consumersbuy f romstoreAwhile50buy f romstoreB.
(i) Write down A’s and B’s payoff/profit in terms of pa and pb
(ii) Find out all Nash equilibria in pure strategies.
(iii) Are all Nash equilibria in (ii) admissible?
 ECON 2112. PROBLEM SET 3

DUE: 13 MARCH, SATURDAY, 5 PM

Please write your name and ID on the first page of your assignment

Exercise 1. [3 points] Airbus and Boeing - A and B hereafter - are two major players in the
market for aircrafts. A is deciding whether to enter a new market. If A stays Out, A receives 0
and B bags $100 million in profit. If A enters, i.e., if A stays In, B can either start a Price war
or Accommodate. If B chooses Accommodate, each receives $30 million. Else, if B chooses
Price war, each loses $10 million (i.e., think of this as −10).
(i) Write down the extensive-form game. [Note: It is sufficient to draw just a game tree
specifying players’ names at the decision nodes, choices in the branches, and payoffs
at the terminal nodes]
(ii) Write down the set of pure strategies available to each player.
(iii) Solve the game using backward induction.
(iv) Now consider the normal-form game where payoffs are as stated above but A and B
are moving simultaneously. Draw the payoff matrix with A being the row player and
B is the column player]
(v) State all the Nash equilibria in pure strategies.
(vi) Are all Nash equilibria in (iv) admissible?

Exercise 2. [3 points] Consider the following perfect information game:

1
L R
1 2
a b c d

1, 1 0, 0 0, 0 1, 1

(i) Solve the game using backwards induction and write down the behavioural strategy
profile and the pure strategy profile (that conforms with backward induction)
(ii) Now consider the normal form representation of the game.
• Specify the set of pure strategies for both players
• Draw the payoff matrix with player 1 as the row player and player 2 as the column
player
• State all Nash equilibria in pure strategies.
• Find the set of admissible equilibria.

1
2

Exercise 3. [3 points]Solve the following perfect information game using backwards induc-
tion:
1
a b
2 2
c d e f
N 1
1 1
2 2 0, 2 g h 1, 1
2
4, 2 3, 0 3, 0
j m
k
1 1
n o 4, 3 p q
N
1 2
1, 2 3 3 2, 5 3, 2

4, 6 1, 3

Exercise 4. [1 point] Recall Traveller’s Dilemma from the Welcome Video. Consider a slight
variation of the game where the airline manager asks Lucy to say a number first and Pete can
go next. Pete will know what Lucy says when he states the price of the gift. Assume both Pete
and Lucy are interested in maximizing just their own payoffs.
• What number will Lucy say? What about Pete?
• What will Lucy and Pete get?
Briefly explain your response to (i) and (ii). [Hint: One way to think about it is that suppose
Lucy says L ∈ {2, 3, ..., 100}. What is Pete’s best response? Anticipating that what will Lucy
say.]
 ECON 2112. PROBLEM SET 1

DUE: 5 PM, 1ST APRIL (THURSDAY)

Instructions
• Please write your name and student ID on the top of assignment
• Please write your tutorial time (e.g. Wednesday 3-4.30) and submit it into the submission box
(which has your tutor’s name). For example, if your tutor is Chris, please do not submit your
assignment in the box meant for DJ’s or Fabio’s students. Details of tutorials and tutor names
are listed under Tutorials in the Moodle website for this course.
• 3 marks will be deducted for assignments without names/ID/tutorial times, submission in the
wrong box, and late submissions (within one hour of submission time). Assignments submit-
ted more than an hour late will not be graded unless the student has received prior approval
(possibly due to exceptional circumstances).
• Note that the deadline for this submission is Thursday.
• While the questions are long as they provide context, the answers (per exercise), written pre-
cisely, are expected to be within a page.

1
2

Exercise 1. Tolls (8 marks)

If Lauren drives to work, she has to use two toll roads - B(ertrand) and C(ournot) - named after the two
firms which built these two toll ways. Let T B be the toll charged by firm B for using Bertrand. Similarly,
let TC be the toll charged by firm C for using Cournot. Lauren is willing to pay at most $12 (in tolls) per
trip. Her demand for trips per week is given by
Q = 12 − T B − TC
Suppose the two firms, B and C, have no other costs. Each firm i(= B, C) chooses T i to maximize its
own profit πi = T i Q.
Part 1

(i) (2 marks) Find T B that maximizes firm B’s profit, πB , for a given TC . Similarly, find TC that
maximizes firm C’s profit for a given T B .
(ii) (2 marks) Suppose both firms set toll simultaneously. Let (T B∗ , TC∗ ) denote the Nash equilibrium
prices. Find (T B∗ , TC∗ ) and profits of firms B and C in Nash equilibrium.
(iii) (2 marks) Suppose firm B chooses T B first and firm C chooses TC afterwards. What toll will
firm B choose?
(iv) (2 marks) Election is in the corner. Opposition leader, Chandra, is proposing a price cap of $3
on tolls. That is, T A ≤ 3, T B ≤ $3. How will it affect your answer to part (ii)? Would firms be
hurt with the price cap? [Treat profits in part (ii) as benchmark for comparison]

Part 2: Not for submission/grading

Price cap turned out to be popular with voters. Chandra won the last election. Almost three years have
passed and another election is coming up. Chandra is up against Alison who is ahead in polls. Only
sticking point for Alison is the price cap. She is proposing to eliminate the price cap as interference with
market does not sit well with her. Alison is trying to convince the electorate that even if the price-cap is
taken off, Bertrand and Cournot would not revert back to higher tolls. Does Alison have a point? [Hint:
Think in terms of repeated games]
 3

Exercise 2. Wages and Employment in a Unionized Firm (6 marks)

* This exercise is based on section 2.1.C in the textbook by Gibbons: Game Theory for Applied
Economists.

Part 1
Consider a model of firm and monopoly union where the union has exclusive control over wages (w)
while the firm has exclusive control over employment (L). Union’s utility function is given by:
U(w, L) = (w − 15)L. (0.1)
The firm’s profit function is given by:
L2
π(w, L) = 55L − − wL, (0.2)
2
L2
where wL is the wage bill and 55L − 2 is the revenue that the firm can earn by employing L workers.

Suppose the timing of the game is as follows: (1) the union makes a wage demand, w; (2) the firm ob-
serves (and accepts) w and then chooses employment, L; (3) payoffs are U(w, L) and π(w, L) as specified
in (0.1) and (0.2).
(i) (2 marks) Find the amount of labour (L) that the firm chooses in stage 2 for a given w. Let
L(w) denote firm’s choice of L given w.
(ii) (2 marks) Suppose union chooses w first, following which the firm chooses L. What w will
union choose? What is the profit-maximizing level of L corresponding to that w (chosen by the
union)?
(iii) Let w∗ and L∗ denote answers to part (ii).
• (1 mark) Find π( w∗ , L∗ ) and U(w∗ , L∗ )
• (1 mark) Are there values of (w, L) such that firm’s profit is strictly higher than π( L∗ , w∗ )
and union’s utility is strictly higher than U(L∗ , w∗ )? Suggest one such value and verify.

Part 2: Not for submission/grading

Consider two extended games.

(i) Game 1: The firm proposes a (w, L) bundle. If the union accepts, the game ends. Else, it
proceeds to the sequential move game described above where the union makes a wage offer
first following which firm chooses employment. Assume that, if indifferent between accepting
a offer and not accepting an offer (i.e., if both options give the same payoff) the union accepts
the offer. What (w, L) will the firm offer?
(ii) Game 2: Now consider the other sequence where the union proposes (w, L). If the firm accepts
union’s proposal, the game ends. Else, it proceeds to the game described in Part 1. If indifferent
between accepting a offer and not accepting an offer (i.e., if both give the same payoff) the firm
accepts the offer. What offer will the union make?
4

Exercise 3. Repeated Games (6 marks)

Remember Little Italy from PS 2? It has 100 consumers all of whom buy either one pizza or nothing.
Each consumer has a reservation price of $15. That is, each consumer is willing to pay at most $15 for
a pizza. There are two pizza stores - A and B - located next to each other. It costs $7 to make a pizza.
Let pa and pb denote the price charged by stores A and B respectively. For simplicity, assume
pi ∈ {$12, $15}
for both i = A, B. Hereafter we ignore the dollar sign. Thus, prices can take just one of the two values -
12 or 15. If pi < p j , all consumers buy pizza from store i. If pi = p j = p then 50 consumers buy from
store A while 50 buy from store B.

(i) (1 mark) Write down the payoff matrix. State the unique Nash equilibrium of the one-shot
game.
(ii) (1 mark) Suppose the game is repeated twice. Is there a subgame perfect equilibrium where
pa = pb = 15 in the first period? Assume that after first period, one-period profits are realised
and each store knows whether or not the other store has charged 15 or 12.
(iii) (1 mark) Suppose the game is infinitely repeated. Consider the following trigger strategy by
player i: Play pi = 15 in the first stage. In the t-th stage, if the outcome of all t − 1 preceding
stages has been (15,15) then play pi = 15; otherwise, play pi = 12. Assume a common discount
factor of δ for both stores A and B. Can the two stores sustain pa = pb = 15 each period if
δ = 32 ? What about δ = 31 ?
(iv) (3 marks) Now add pi = 9 to the set of possible prices. That is, suppose
pi ∈ {9, 12, 15}
How does the inclusion of this new strategy pi = 9 affect your findings. In particular,
• Are there any additional pure-strategy Nash equilibrium in the one-shot game?
• Does you answer to part (ii) change with inclusion of this additional strategy?
• Does your answer to part (iii) change? [Assume δ = 13 ]
 ECON 2112. PROBLEM SET 4

SOLUTION TO PROBLEM 2 WITH BRIEF EXPLANATIONS

1
2

Wages and Employment in a Unionized Firm (3 marks)

* This exercise is based on section 2.1.C in the textbook by Gibbons: Game Theory for Applied
Economists.

Consider a model of firm and monopoly union where the union has exclusive control over wages (w)
while the firm has exclusive control over employment (L). Union’s utility function is given by:

U(w, L) = (w − 15)L. (0.1)

The firm’s profit function is given by:

L2
π( w, L) = 55L − − wL, (0.2)
2
L2
where wL is the wage bill and 55L − 2 is the revenue that the firm can earn by employing L workers.

Suppose the timing of the game is as follows: (1) the union makes a wage demand, w; (2) the firm
observes (and accepts) w and then chooses employment, L; (3) payoffs are U(w, L) and π( w, L) as
(i) (1 mark) Find the amount of labour (L) that the firm chooses in stage 2 for a given w. Let L(w)
denote firm’s choice of L given w.

∂π( w,L)
Answer: Set ∂L = 55−L−w = 0 and solve for L in terms of w. This gives L = 55−w ≡ L(w).
∂2 π w,L)
[Qualifiers - (a) Check that second order-condition is satisfied, i.e. ∂L( 2 = −1 < 0; (b) to be
very precise, L(w) = max{55 − w, 0} which takes into account that L = 0 if w ≥ 55.]

(ii) (2 marks) Suppose union chooses w first, following which the firm chooses L. What w will
union choose? What is the profit-maximizing level of L corresponding to that w (chosen by the
union)?

Answer: Substituting L = L(w) in (0.1) gives:

U = (w − 15)(55 − w).

dw = (55 − w) − (w − 15) = 0 and rearranging yields: w = = 35. Union will

15+55
Setting dU 2
choose w = 35. Profit-maximizing level of L corresponding to w = 35 is L = 55 − 35 = 20.
2
[Qualifiers - Once again worth checking that second order condition is satisfied, i.e., ddwU2 =
−2 < 0]

(iii) (2 marks) Let w∗ and L∗ denote answers to part (ii).

• Find π( w∗ , L∗ ) and U(w∗ , L∗ )

• Are there values of (w, L) such that firm’s profit is strictly higher than π( L∗ , w∗ ) and union’s
utility is strictly higher than U(L∗ , w∗ )? Suggest one and verify.

Answer: (i) We have

202
π( w∗ , L∗ ) = π( 35, 20) = 55 ∗ 20 − − 35 ∗ 20 = 200
2
U(w∗ , L∗ ) = U(35, 20) = (35 − 15) ∗ 20 = 400
 3

(ii) Check with L = 40 and w = 28. Check that

402
π(28, 40) = 55 ∗ 40 − − 28 ∗ 40 = 280 > 200
2
U(28, 40) = (28 − 15) ∗ 40 = 520 > 400

Ungraded parts:

Union chooses (w, L) to maximize

(w − 15)L
subject to
L2
π(w, L) = 55L − − wL ≥ 200
2
The constraint implies
L2
wL ≤ 55L − − 200.
2
Substituting this in union’s objective function, union’s problem reduces to choosing L maximising
L2
(55 − 15)L − − 200.
2
Check that L∗ = 40. Now substitute L = 40 in firm’s profit expression and solve for w that ensures
π(w, 40) ≥ 200
Setting the profit constraint as equality and solving for w yields w = 30.

Now consider the other side. Firm chooses (w, L) to maximize

L2
π(w, L) = 55L − − wL
2
subject to
(w − 15)L ≥ 400
The constraint implies
wL ≥ 15L + 400.
Setting this constraint as equality, substituting that into the maximand, and simplifying subsequently
reduces the firm’s problem to the following: Choose L to maximise
L2
55L − − 15L − 400
2
Once again, check that answer is L∗ = 40. Union’s constraint, i.e.,
(w − 15)L ≥ 400
is satisfied at L = 40 when w = 25.

Note that L∗ is the same for both constrained optimization problems. What is different is the level
of w. Note, though, wages in both optimization problems (in the ungraded part) are lower than union’s
original demand of 35.
 ECON 2112. PROBLEM SET 5

DUE: 5 PM, 10TH APRIL (SATURDAY)

Instructions
• Please write your name and student ID on the top of assignment
• Please write your tutorial time (e.g. Wednesday 3-4.30) and submit it into the submission box
(which has your tutor’s name). For example, if your tutor is Chris, please do not submit your
assignment in the box meant for DJ’s or Fabio’s students. Details of tutorials and tutor names
are listed under Tutorials in the Moodle website for this course.
• 3 marks will be deducted for assignments without names/ID/tutorial times, submission in the
wrong box, and late submissions (within one hour of submission time). Assignments submit-
ted more than an hour late will not be graded unless the student has received prior approval
(possibly due to exceptional circumstances).
• Note that the deadline for this submission is 10 April, Saturday, 5 pm.

1
2

Exercise 1. Question 1 (Fight or Cave)

There are two players, A and B. Each player i ∈ {A, B} can be of one of two types:
ti ∈ {1, 2} .
The probability that a player is of type 2 equals p. When A and B meet, each can decide to fight or cave.
If both players fight, then player i gets payoff
ti
−c
ti + t j
where j , i and c>0. If player i fights, but player j does not, then player i gets payoff 1, while player j
gets payoff 0. If both players do not fight, each gets payoff 12 .

(a) (1 mark) Draw the Bayesian normal form representation of this game.

(b) (1 mark) A strategy in a static Bayesian game is a function that specifies an action for each type of
a player. Write down all the possible strategies for player i.

(c)(2 marks) Assume that A plays fight if tA = 2 and cave otherwise.

(i) If B is of type 2, what should B do?
(ii) If B is of type 1, what should B do?

(d) (2 marks) Is there a Bayesian Nash equilibrium in which each player fights if and only if she is of
type 2? If so, what is the equilibrium probability of a fight?

(e) (2 marks) Assume that A never fights.

(i) If B is of type 2, what should B do?
(ii) If B is of type 1, what should B do?

(f) (2 marks) Is there a Bayesian Nash equilibrium in which no player ever fights?
 Problem Set 6
ECON 2112
Deadline: 17th April, 5 pm

• Please write your name and student ID on the top of assignment

• Please write your tutorial time (e.g., Wednesday 3-4.30) and submit it into the
submission box (which has your tutor’s name). For example, if your tutor is Chris,
please do not submit your assignment in the box meant for DJ’s or Fabio’s students.
Details of tutorials and tutor names are listed under Tutorials in the Moodle website
for this course.
• 3 marks will be deducted for assignments without names/ID/tutorial times,
submission in the wrong box, and late submissions (within one hour of submission
time).
• Assignments submitted more than an hour late will not be graded unless the student
has received prior approval (possibly due to exceptional circumstances).
Question
There are two players called 1 and 2. Player 1 can be of two types t ∈ {0,1} with Pr (t=1) = π
∈ (0,1). The actions and payoffs of the game are given by:

left right
up 0,4 1,1
down 1,2 t,4

where the row player is player 1.

We will use the following notation:
• σ1(t): probability that player 1 plays up if she is of type t;
• σ2: probability that player 2 plays left.

Part I (2 marks)

Suppose π = 0.5.

Is (σ1(0), σ1(1), σ2) = (0,1,0) a Bayes-Nash equilibrium?

[Hint: To prove something is a BNE you have to check no player-type has an incentive to
deviate from the proposed strategy profile. To prove something is not a BNE, you need to
check just one of the three – type-0 player 1, type-1 player 1, or player 2 has an incentive to
deviate]

Part II (8 marks)

We want to know whether and when it is possible that in a Bayes Nash equilibrium player 1
mixes between up and down whenever she is of type t = 0, i.e. σ1(0) ∈ (0,1).
We therefore proceed to construct such an equilibrium and then verify for which values of π
this equilibrium exists.
At the end of the exercise, you should complete the following “Proposition”

Proposition 1. If π ∈ (… , …), then there exists a Bayes Nash equilibrium in

which player 1 mixes between up and down whenever she is of type t=0 ,
i.e. σ1(0) ∈ (0,1). In this equilibrium σ1 (0) = …; σ1(1) = …; σ2 = … … … .

1. (1 mark) If type-0 player 1 is mixing, what condition must be satisfied in this

equilibrium? (Hint: if I am mixing then it means that I am …)
2. (1 mark) Using the condition derived in part 1, you should be able to find player 2 's
equilibrium strategy σ2. What is it?

3. (2 marks) Using your answers to parts 1 and 2, we can immediately conclude that in
this equilibrium type-1 player 1 must play...? (Hint: remember to state your answer as
a value for σ1(1))

4. (3 marks) Now you should be able to find σ1(0). What is it? [Hint: the answer is a
formula containing π. Notice that it is easy to mess up signs when calculating σ1(0),
so be careful and double-check your math.

5. (1 mark) You now have a complete profile of strategies given by σ1(0), σ1(1), σ2. But
you can notice that for some values of π it is not true that σ1(0) ∈ (0,1). Find the
values of π for which σ1(0) ∈ (0,1).
 ECON 2112. PROBLEM SET 1: SOLUTIONS

Exercise 1.

(i) Each player has two strategies, let us call them Stag and Hare.

Stag Hare
(ii)
Stag 2, 2 0, 1
Hare 1, 0 1, 1

(iii) The two pure strategy Nash equilibria are (Stag, Stag) and (Hare, Hare).

Exercise 2: See worked out solutions

Exercise 3. See the lecture notes for the appropriate definitions.

Exercise 4: See the next page

Exercise 5.

L M R
T 0, 0 0, 0 0, 0
(i)
C 1, 0 0, 1 1, 0
B 0, 0 1, -1 0, 0

L M R
T 1, 1 0, 0 0, 0
(ii)
C 0, 0 1, 1 0, 0
B 0, 0 0, 0 1, 1

L M R
T 0, 0 0, 0 0, 0
(iii)
C 0, 0 0, 0 0, 0
B 0, 0 0, 0 0, 0

1
2

Exercise 2.
 ECON 2112. PROBLEM SET 2

DUE: SATURDAY 5 PM, 6TH MARCH

*Please write your name and ID on the first page of your assignment

Exercise 1. Finding Nash equilibria

Find all pure and mixed strategy Nash equilibria of the following games.
(i)

L R
T 0, 0 1, 3
B 3, 1 0, 0
(ii)

L R
T 9, −9 7, −7
B 6, −6 8, −8
(iii)

L C R
T 4, 3 2, 7 0, 4
B 5, 5 5, −1 −4, −2

(i) (T,R) (B, L), (1/4 T + 3/4 B, 1/4 L + 3/4 R)

(ii) (1/2 T + 1/2 B, 1/4 L + 3/4 R)

(iii) (B,L). (There is no Nash equilibrium in mixed strategies.)

1
2

Exercise 2. Voting
Voters in Suburbia are choosing between two candidates, L and R. L is proposing new infras-
tructure funded with new taxes. Candidate R is in favour of preserving the status quo, i.e., no
new taxes and no new spending. Consider three voters - Alison, Beth, and Chandra - whose
preferences and payoff are as follows.
• Alison strictly prefers L to R. She gets a payoff of 2 if L wins but 0 otherwise.
• Beth strictly prefers R to L. She gets a payoff of 2 if R wins but 0 otherwise.
• Chandra is not interested in electoral politics. Nonetheless voting is compulsory in
Suburbia so he always votes. He gets 0 if the candidate he votes for wins. Otherwise,
his payoff is -1.

All three individuals - Alison, Beth, and Chandra - vote. Candidates receiving majority of the
votes wins.
(i) Write down the 3-person normal form game where Alison, Beth, and Chandra are three
players and each of them has two strategies - vote for L or vote for R. As described
above, their payoffs depend on the outcome of the election.
(ii) Find out all Nash equilibria in pure strategies.
(iii) Suppose we are interested in Nash equilibria where neither strictly nor weakly domi-
nated strategies are played. A Nash equilibrium satisfying this property is called ad-
missible. Are all Nash equilibria in (ii) admissible?
 3

Exercise 3. Entry into new markets

Hunter and Melissa are two competing coffee growers who are exploring new markets to sell
their product. There are two potential markets - Large and Small - referred to as L and S
hereafter. Entry costs are large so each of them can enter only one of the two markets.
• If one of them enters L while the other one enters S, then one who enters L gets 36, while
the other one gets 9.
• If both enter L, each gets 36α.
• If both enter S, each gets 9α.
where 0 ≤ α ≤ 21 .

(i) The game described above has at least one Nash equilibrium in pure strategies irrespective
of the value of α - Do you agree?
(ii) The game described above has at least one Nash equilibrium in strictly mixed strategies
irrespective of the value of α - Do you agree? [Note: By strictly mixed strategies we refer
to strategies of the form (xL + (1 − x)S , yL + (1 − y)S ) where 0 < x < 1 and 0 < y < 1.]
4

Exercise 4. Price competition

There are 100 consumers in the suburb Little Italy all of whom buy either one medium pizza
or nothing. Each consumer has a reservation price of $15. That is, each consumer is willing
to pay at most $15 for a medium pizza. There are two pizza stores - A and B - located next to
each other. It costs $7 to make a medium pizza. The price of pizza is in whole dollars (e.g., $8,
$7, but not $6.99 or $8.50)
Let pa and pb denote the price charged by stores A and B respectively. For simplicity, assume
pi ∈ {1, 2, ..., 20} for both i = A, B. If pi < p j , all consumers buy pizza from store i as long as
pi ≤ $15. If pi = p j = p, and p ≤15, then50consumersbuy f romstoreAwhile50buy f romstoreB.
(i) Write down A’s and B’s payoff/profit in terms of pa and pb
(ii) Find out all Nash equilibria in pure strategies.
(iii) Are all Nash equilibria in (ii) admissible?
 0

- o �� r� = 1
1! .(.._ o � r ll f e-
o.. L_ 1"
-, � Tl � (� I �) > Tl � Cr IA ) � ) ti r C. l
€ in . . 1-�)
I

(� 1 CZ) Vj � 0t

r k - 1- -"' �1- � 1- p ,,_ � �

� -<-+.:--J(Av-< { h
7'Tc-. (1 , p b ) =- 0 -Y-v'

tr� [<JI �!,) -{O

+ \H_ �� �b >�
=--J Tl"' ( 'il / l'b) '> TT,,__ (-:i hJ �
J h,) 'ii
--::. \lc_(7-/� \;i} f,-... p�<g
\:, "' -:: f - v- � k} � ,,___,)--LJ - -=-) fl- 1 1) lA..J
r;__&_ �<;¾-�.

Pb > 4
l D o
TL� ct, ) r b) 0 0
' 0

r� :: < -
( 9' .,_ ("l - 1,V

N �� F • � '7 � - � � >--4.,-w_
� � (� / �) � ( 1, �) . � � �/"n't•�
 ECON 2112. PROBLEM SET 3

DUE: 13 MARCH, SATURDAY, 5 PM

Please write your name and ID on the first page of your assignment

Exercise 1. [3 points] Airbus and Boeing - A and B hereafter - are two major players in the
market for aircrafts. A is deciding whether to enter a new market. If A stays Out, A receives 0
and B bags $100 million in profit. If A enters, i.e., if A stays In, B can either start a Price war
or Accommodate. If B chooses Accommodate, each receives $30 million. Else, if B chooses
Price war, each loses $10 million (i.e., think of this as -10).

(i) Write down the extensive-form game. [Note: It is sufficient to draw just a game tree
specifying players' names at the decision nodes, choices in the branches, and payoffs
at the terminal nodes]
(ii) Write down the set of pure strategies available to each player.
(iii) Solve the game using backward induction.
(iv) Now consider the normal-form game where payoffs are as stated above but A and B
are moving simultaneously. Draw the payoff matrix with A being the row player and
B is the column player]
(v) State all the Nash equilibria in pure strategies.
(vi) Are all Nash equilibria in (iv) admissible?

A (_t')

$ A := t _I vi I
O v-tJ
(o 1 L00
1_A-�/
P�u.- Wwe-l

(}h) A-�)
� � (007-\�

A � r L- Crr> '7 o)
rt�-
Iv> i,_;,,i,._ � �
�)-

� l> �--
2

Exercise 2. [3 points] Consider the following perfect information game:

1

1, 1 0,0 0,0 1, 1

(i) Solve the game using backwards induction and write down the behavioural strategy
profile and the pure strategy profile (that conforms with backward induction)
(ii) Now consider the 1101mal form representation of the game.
� • Specify the set of pure strategies for both players
-J • Draw the payoff matrix with player 1 as the row player and player 2 as the column
player
--) • State all Nash equilibria in pure strategies.
-J •

u
i
,
2

Pa:cts-fi-H�ee-att:ached figure-2 �tne assoc1atedaescnphon there.

�
��-�Ak,;vJ-�� �

� �� ���-�
J_,)
GL -t- ( \ ---x ) \2-;
(;....

0 <'X- -
.....
� \

0 t� �AAl- rr �µ,
(_LO\..; J-)

(R-A- ,1 J_)
 ECON 2112. PROBLEM SET 1

DUE: 5 PM, 2ND APRIL (FRIDAY)

*Please write your name and ID on the top of assignment

Exercise 1. Setting Tolls (3 marks)

Parts (i) and (ii):

Firm i’s profit is given by

πi = T i (12 − T i − T j )
∂πi
where i, j ∈ {B, C}, j , i. Setting ∂T i = 0 for i = B, C and rearranging we get:
12 − TC 12 − T B
TB = , TC =
2 2
which respectively are the profit-maximizing values B and C.

[Note: We do not need to take points off for this. But technically speaking, one needs to be more precise
- and something to mention in tutorials. First, we need to check that second order condition is satisfied:
∂2 πi
∂T 2
< 0 Second, the best response function is
i

12 − T j
T i = max{ , 0}
2
which tells us what happens when T j > 12. ]

Solving the two best response functions yields:

T A∗ = T B∗ = 4.

Part (iii) Firm B chooses T B to maximize

T B (12 − T B − TC )

where TC = 12−T B
2 . Substituting the constraint in the maximand, differentiating, and setting the resulting
expression to zero yields:
TB = 6

Part (iv) The best-response functions of Firms B and C respectively are

12 − TC 12 − T B
T B = max{ , 3} TC = max{ , 3}
2 2
Substituting TC = 3 in B’s best response function, and substituting T B = 3 in C’s best-response function
we get T B = 3, TC = 3. Thus (T B , TC ) = (3, 3) is a Nash equilibrium.
1
2

[Note: It might be easier to draw a diagram here. For example, if you are drawing best response function
for firm B, it woud be T B = 12−T 2
C
if TC ≥ 6 and T B = 3 if TC < 6. Similarly draw it for firm C. The
intersection will be in the vertical segments, i.e. where T B = 3 if TC = 3. Price-cap need to feature as
NE, i.e. sticking to price-cap is a best response when rival firm sticks to price-cap.

Non-graded part

Suppose T B = TC = 3. Then, QB = QC = 12 − 3 − 3 = 6 and each firm’s profit is 3 ∗ 6 = 18. If

TC = 3, firm B’s profit, given by
T B (12 − 3 − T B )
is maximized at T B = 4.5. Its profit at (T B , TC ) = (4.5, 3) is (4.5)2 = 20.25. Nash equilibrium profits for
each firm, as stated in part (ii) is 16.
Suppose the game is infinitely repeated and each firm plays the following: Play T B = 3 in the first
stage. In the t − th stage, if the outcome of all t − 1 preceding stages has been (3, 3), then play T i = 3;
otherwise, play T i = 4.
Low tolls, i.e. (T B , TC ) = (3, 3) can be sustained if there is no incentive to deviate, i.e.
18 16δ
≥ 20.25 +
1−δ 1−δ
or δ ≥ 17
9
= 0.529. Alison have a point if discount factor is high enough.
ECON2112 Game Theory and Business Strategy
Solutions to PS 5

Chris Teh

April 19, 2021

Chris Teh April 19, 2021 1 / 10

The Model

There are two players, A and B. Each player i ∈ {A, B} can be one of two
types, ti ∈ {1, 2}. Pr (ti = 2) = p ∈ (0, 1). When A and B meet, they can
decide whether to fight or cave (Ai = {F , C }). If both players fight, then
player i gets a payoff of
ti
−c
ti + tj
Where j 6= i and c > 0. If player i fights, but player j does not, then
player i gets payoff 1, while player j gets payoff of 0. If both do not fight,
they get payoff of 1/2.

Chris Teh April 19, 2021 2 / 10

The Model

1a) Draw the Bayesian normal form representation of this game

Player B

 Fight  Cave
tA
Fight tA +tB − c, tAt+t
B
−c (1, 0)
Player A B
 
1 1
Cave (0, 1) 2, 2

Note: the constant c > 0 and prior probabilities p ∈ (0, 1) play a role in
determining best-responses.

Chris Teh April 19, 2021 3 / 10

Strategies (with math)

1b) Write down all possible strategies for player i

Recall: A strategy is a mapping from types to actions, i.e., si : Ti → Ai .

My notation: Let si = (a1 , a2 ) represent a strategy where type 1 of player

i plays action a1 , and type 2 of player i plays action a2 .

Since Ti has two elements, each player’s strategy set will have four
elements. That is,
 
Si = (F , C ), (F , F ), (C , F ), (C , C )

Chris Teh April 19, 2021 4 / 10

Strategies (in english)

For each player i, his strategies are:

Type 1 plays fight, Type 2 plays fight (aggressive)
Type 1 plays fight, Type 2 plays cave (dumb)
Type 1 plays cave, Type 2 plays fight (smart)
Type 1 plays cave, Type 2 plays cave (cowardly)

Chris Teh April 19, 2021 5 / 10

Suppose A plays fight (cave) if tA = 2 (tA = 1). If B is of type 2, what should he

do?

Player B
Fight Cave
Fight ( tAt+t
A
− c, tB
tA +tB
− c) (1, 0)
Player A B

Cave (0, 1) ( 12 , 12 )

Fix say sA = (C , F ) and tB = 2. Then

E[uB (F , sA |t)|tB = 2] = 1 − cp − p/2

E[uB (F , sA |t)|tB = 2] = 1/2 − p/2

Notice: B weakly prefers F when

1 − cp − 1/2p ≥ 1/2 − p/2 ⇐⇒ 1/2p ≥ c

Chris Teh April 19, 2021 6 / 10

Suppose A plays fight (cave) if tA = 2 (tA = 1). If B is of type 1, what should he

do?

Player B
Fight Cave
Fight ( tAt+t
A
− c, tB
tA +tB
− c) (1, 0)
Player A B

Cave (0, 1) ( 12 , 12 )

Fix say sA = (C , F ) and tB = 1. Then

E[uB (F , sA |t)|tB = 1] = 1 − cp − 2p/3

E[uB (C , sA |t)|tB = 1] = 1/2 − p/2

Notice: B weakly prefers F when

1 − cp − p/2 ≥ 1/2 − p/2 ⇐⇒ 1/2p − 1/6 ≤ c

Chris Teh April 19, 2021 7 / 10

Is there a Bayes-Nash equilibrium in which each player fights if and only if she is of

type 2?

Fight is a best-response to (Cave, Fight) for type 2 if and only if

1
c≤
2p

Cave is a best-response to (Cave, Fight) for type 1 if and only if

3−p 1 1
c≥ ≡ −
6p 2p 6

Thus, (Cave, Fight) is a best-response to (Cave, Fight) when

1 1 1
− ≤c ≤
2p 6 2p

Probability of fight = Pr (t1 = 2, t2 = 2) = p 2

Chris Teh April 19, 2021 8 / 10

Suppose A plays (Cave, Cave). What should a type 2 B do?

Player B
Fight Cave
Fight ( tAt+t
A
− c, tB
tA +tB
− c) (1, 0)
Player A B

Cave (0, 1) ( 12 , 12 )

Notice: Fix sA = (C , C ). Player B’s type does not affect his payoffs
from playing F or C . Thus, the best-response of type 2 B to (C , C )
is the same as that of type 1 B.
Fixing sA = (C , C ), B’s payoff from playing F is 1. This is strictly
greater than the payoff from C of 1/2.
Thus, B will play F (always) in response to (C , C ).

Chris Teh April 19, 2021 9 / 10

Is there a Bayes-Nash equilibrium in which no player ever fights?

Here, the best response of type 2 to (Cave, Cave) is strictly to Fight.

Also, the best-response of type 1 to (Cave, Cave) is strictly to Fight.

Therefore, there does not exist a Bayes-Nash equilibrium in which both

players play (Cave, Cave) (always deviate to F ).

Chris Teh April 19, 2021 10 / 10

The traveller’s dilemma (Kaushik Basu, Scientific American, June 2007)

Returning from a remote island, Lucy and Pete, find that the airline has damaged the identical
antiques that each had purchased. An airline manager says that he is happy to compensate
them but is handicapped by being clueless about the value of these strange objects. Simply
asking the travelers for the price is hopeless, he figures, for they will inflate it. Instead he asks
each of them to write down the price of the antique as any dollar integer between 2 and 100
without conferring together. If both write the same number, he will take that to be the true
price, and he will pay each of them that amount. But if they write different numbers, he will
assume that the lower one is the actual price and that the person writing the higher number is
cheating. In that case, he will pay both of them the lower number along with a bonus and a
penalty—the person who wrote the lower number will get $2 more as a reward for honesty and
the one who wrote the higher number will get $2 less as a punishment. For instance, if Lucy
writes 46 and Pete writes 100, Lucy will get $48 and Pete will get $44. What numbers will
Lucy and Pete write? What number would you write? 6
 ECON 2112. PROBLEM SET 1, EXERCISE 3

Exercise 3 (Strategic Subsidies)Airbus (a European firm) and Boeing (an US firm) are competitors in the market
for aircrafts. The demand for aircrafts has gone down significantly in recent times - in fact, so much so that only
one player can profitably operate in the market. If both stay in the market, each loses $5 million. If both exit the
market each gets 0, where as if one stays and one exits, the one who exits gets 0 while the one who stays bags
$100 million.
(i) Identify the set of pure strategies for each player.
(ii) Write down the normal form game and the associated payoff matrix.
(iii) Compute the set of pure strategy Nash equilibria.
Now suppose the US government announces and credibly commits to $6 million subsidy for Boeing, contingent
on Boeing not exiting. That is, Boeing will receive $6 million as long as it remains in the market[note: this
means $6 million gets added to Boeing’s profit/loss figures stated above provided it stays]. How does this affect
your answers to (ii) and (iii)?
Solution: Suppose Airbus (A) is the row player and Boeing (B) is the column player. Let S i denote the strategy
set of player where i = A, B. From the description it follows that
S i = {S tay, Exit}
S 1 × S 2 = {(S tay, S tay), (S tay, Exit), (Exit, S tay), (Exit, Exit)}
For simplicity, hereafter, we use S and E to denote Stay and Exit respectively. Let ui (si , s j ) denote player i’s
payoff corresponding to the strategy profile (s1 , s2 ) ∈ S 1 × S 2 .
uA (S , S ) = −5, uA (S , E) = 100, uA (E, E) = 0, uA (E, S ) = 0;
uB (S , S ) = −5, uB (S , E) = 0, uB (E, E) = 0, uA (E, S ) = 100;
The payoff matrix is as follows:

S E
S −5, −5 100, 0
E 0, 100 0, 0

There are two Nash equilibria in pure strategies: (S , E) and (E, S ).

Now consider the game with subsidies. Payoffs for Boeing increase by $6 million but only if it stays in the
market. Thus
uB(S , S ) = −5 + 6 = 1, uB(E , S) = 100 + 6 = 106.
All other payoffs are same as before. The new payoff matrix is as follows:

S E
S -5, 1 100, 0
E 0, 106 0, 0

There is a unique Nash equilibrium: (E , S). Airbus staying in the market is no longer a NE.