Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: THEend8_
ECON 20005: Competition and Strategy
Tutorial Twelve (Sample Final) Solutions
Note: These are solution sketches that do not necessarily show all the work/intermediate
steps as required for full marks on the final. It is your task to clearly and succinctly
illustrate your knowledge on the final exam. When in doubt, show your work!
Question 1 (15 marks)
a. Find all pure strategy Nash Equilibria of the following two simultaneous- move games. If
there is no pure strategy Nash Equilibrium in one of these games, you should explain why
not.
(i)
P2
A B C
P1
a 9, 3 7, 4 3, 1
b 3, 6 11, 2 1, 4
c 6, 4 2, 3 8, 5
(ii)
P2
p q r
P1
m 2,−2 5,−5 13,−13
n 6,−6 5.6,−5.6 10.5,−10.5
o 10,−10 3,−3 −2, 2
(i) There is one pure strategy Nash Equilibrium: (c, C)
(ii) There is one pure strategy Nash Equilibrium: (n, q)
b. Consider the following sequential-move game:
BA
1
d
1, 1
c
3, 2
2
d
2, 3
c
1, 1
2
(i) Find all the pure-strategy Nash Equilibria to this game.
(iii) If a Nash Equilibrium is not Sub-game Perfect, explain why not.
(i) There pure-strategy NE to the game are (A, cc), (A, cd), (B, dd) (easiest to find these
using the normal form of the game)
(ii) The Nash Equilibira (A, cc) and (B, dd) are not sub-game perfect. They require player
2 to play either d after player 1 having played A, or c after player 1 having played B, both
of which would not be optimal for player 2 if the corresponding sub-game were reached.
Page 1 of 5
c. Consider the following game played between two players. Each player chooses 1, 3, 5 or
7 to add to a cumulative total. The cumulative total starts at 0. The players take it
in turns to choose one of these numbers, and they are able to view the cumulative total
at their turn. The player who causes the cumulative total to equal or exceed 66 is the
loser. Consider the player who starts. Explain whether she wins or loses in the Sub-game
Perfect Nash Equilibrium. Explain her equilibrium strategy.
(i) The first moving player wins in the sub-game perfect Nash Equilibrium.
(ii) There are various strategies that ensure her victory. One is to choose 1 at her first
move and then to ensure that the cumulative total is at 9, 17, 25, 33, 41, 49, 57 and 65
after her subsequent moves, such that her opponent has to cause the cumulative total to
equal or exceed 66. Another is to choose 1 at any move.
Question 2 (15 marks)
This is a penalty kick game in which the goalie has a third option (the option of remaining in the
middle and not reacting until the very last moment). We still assume that it is a simultaneous-
move game. The game table looks as follows:
Goalie
L M R
Kicker
L 30, 70 70, 30 90, 10
R 90, 10 70, 30 30, 70
a. Assume the kicker plays L with probability pL and R with probability 1 − pL , where
0 ≤ pL ≤ 1. Derive the goalies expected payoff for each of his pure strategies, and show
what the goalies best response is to any possible pL.
The goalies expected payoff is 70pL + 10(1 − pL) = 10 + 60pL when playing L, 30 when
playing M , and 10pL + 70(1 − pL) = 70− 60pL when playing R.
His best response is playing L if pL ≥ 0.5, and R if pL ≤ 0.5. (Playing M is never a best
response.)
b. Assume the goalie plays L with probability qL, R with probability qR and M with prob-
ability 1 − qL − qR,where qL ≥ 0,qR ≥ 0 and qL + qR ≤ 1. Derive the kickers expected
payoff for each of his pure strategies, and show what the kickers best response is to any
possible combination of qL and qR.
The kickers expected payoff is 30qL + 70(1 − qL − qR) + 90qR = 70 − 40qL + 20qR when
playing L,and 90qL + 70(1 − qL − qR) + 30qR = 70 + 20qL − 40qR when playing R.
His best response is L if qL ≤ qR,and R if qL ≥ qR.
c. Derive the Nash Equilibrium in mixed strategies.
It follows from parts (a) and (b) that the Nash Equilibrium in mixed strategies is (pL; qL, qR) =
(0.5; 0.5, 0.5)
Question 3 (25 marks)
Consider an industry with two firms Holden and Toyota that face an inverse demand function
P (Q) = 200−Q, where Q is the total quantity produced. Both firms produce at marginal costs
of 20 (and have no fixed costs).
Page 2 of 5
a. Assume that the two firms compete a la Cournot, i.e., that they simultaneously choose
the (continuous) quantity to produce. Derive the Nash equilibrium quantities that the
firms produce as well as their equilibrium profits. (Remember to show your work.)
Holden chooses the quantity qH that maximizes its profit
πH(qH) = (200 − qH − qT − 20)qH = (180 − qH − qT )qH
given qT . Differentiating with respect to qH gives the first order condition
0 = 180− 2qH − qT .
Rearranging gives Holdens best response qH =
180−qT
2
. Similarly, Toyotas best response
is qT =
180−qH
2
.
Inserting Holdens best response into Toyotas best response implies
qT =
180− 180−qT
2
2
⇒ q∗T = 60
Consequently, q∗H = 180− q
∗
T = 180− 60 = 60.
Equilibrium profits are π∗H = π
∗
T = (200 − 2× 60 − 20) × 60 = 60
2 or 3600.
b. Assume that the two firms compete a la Bertrand, i.e., that they simultaneously choose
their (continuous) prices, that consumers buy from the firm that sets the lower price, and
that consumers buy the same quantity from both firms if both firms set the same price.
Find the unique Nash equilibrium in pure strategies and the corresponding equilibrium
profits. Also prove that there cannot exist any other Nash equilibrium.
In the unique Nash equilibrium (NE) prices are p∗H = p
∗
T = 20.
Equilibrium profits are π∗H = π
∗
T = 0.
There cannot exist a NE with min{pH , pT } < 20, as firms making negative profits would
have an incentive to deviate. There cannot exist a NE in which min{pH , pT } > 20, as at
least one firm would have an incentive to deviate and to set a slightly lower price than its
competitor. There can also not exist a NE in which min{pH , pT } = 20 < max{pH , pT },
as the firm setting the lower price would have an incentive to deviate and to increase its
price. Hence the above NE must be unique.
c. Now assume that the government subsidizes Holden with 30 per car that they sell. (In
other words, Holden receives from the government 30 per car that they sell.) Again
assuming Cournot competition, derive the equilibrium quantities and profits of the two
firms.
Now Holden chooses the quantity qH that maximizes its profit
πH(qH) = (200 − qH − qT − 20 + 30)qH = (210 − qH − qT )qH
which leads to its best response qH =
210−qT
2
. Toyotas best response is still qT =
180−qH
2
.
Inserting Holdens best response into Toyotas best response implies
qT =
180− 210−qT
2
2
⇒ q∗T = 50
Consequently, q∗H =
210−q∗
T
2
= 210−50
2
= 80
Equilibrium profits are π∗H = (200 − 80 − 50 − 20 + 30) × 80 = 80
2 = 6400 and π∗T =
(200 − 80− 50− 20) × 50 = 502 = 2500.
Page 3 of 5
d. Now assume again Bertrand competition, and describe one Nash equilibrium in pure
strategies for the case in which the government subsidizes Holden with 30 per car that
they sell. Also derive the firms equilibrium profits.
There are two NE: p∗H = 20 and p
∗
T = 20 + ǫ,or p
∗
H = 20 − ǫ and p
∗
T = 20, with ǫ being
arbitrarily small.
To derive the equilibrium profit of Holden, note that P (Q) = 200 − Q implies Q(P ) =
200 − P . Hence π∗H(200 − 20)(20 − 20 + 30) = 5400, while π
∗
T = 0.
Question 4 (25 marks)
Suppose that a society consists of only two players and that both of them drive a car. Each
player must decide whether to put a pollution control device on their car. Installing such a
device on a car costs $100 and provides a benefit that is worth $70 to each player (both players
benefit since they both breathe cleaner air). If both players install the device, the air is twice
as clean, providing a benefit of $140.
Each player has two possible actions: “Install” or “Not Install.” Assume for the moment
that this game is a simultaneousmove oneshot game.
a. Construct the game table and find all pure strategy Nash Equilibria.
Game table:
Player 2
Install Not
Player 1
Install 40, 40 −30, 70
Not 70,−30 0, 0
The unique Nash equilibrium is (Not,Not).
b. Is this game a Prisoners Dilemma Game? Why/Why not?
It is a Prisoners Dilemma Game as each player has a dominant strategy, and the dominance
solution is worse for both players than one of the non-equilibrium outcomes.
c. Assume for the remainder of the question that the game is played infinitely often, that
both players have a discount factor of δ < 1, and that the pollution device only lasts for
one year so that the two players must decide at the start of each year whether to install
the device.
What infinite payoff will each player receive if they both play a grim strategy and co-
operate forever (by installing the device)?
If they both play a grim strategy and cooperate forever, their infinite payoff is 40
1−δ
.
d. What infinite payoff will a player receive by cheating in this period (and not installing the
device) if the opponent plays a grim strategy?
If the opponent plays a grim strategy, the infinite payoff when cheating in this period is
70.
e. What values of δ sustain the collusive outcome of this game? Provide the intuition for
your solution.
Page 4 of 5
The collusive outcome is sustained if and only if
40δ
1− δ
≥ 70⇒ 40 ≥ 70(1 − δ)⇒ 70δ ≥ 30⇒ δ ≥
3
7
f. Now assume that in each period one of the two players is involved in a deadly car accident
with probability 25%. For what values of δ can the collusive outcome be sustained in this
case?
The effective discount factor becomes 3δ
4
. Hence the collusive outcome is sustained if and
only if
3δ
4
≥
3
7
⇒ δ ≥
4
7
Question 5 (20 marks)
Consider a charitable donations game simultaneously played by N = 100 citizens. Each citizen
must decide whether to donate to $11 charity or not. If n citizens donate to the charity, then
each of the N citizens receives a private benefit worth $n2 regardless of whether they donated
or not (since they feel good about charity donations in general). If an individual donates, they
suffer a private financial cost of $11. Throughout, n must be a whole number from 0 to 100.
a. Find all Nash Equilibria number of donors, nne.
Given n donors, if citizen n + 1 switches from shirking to participating they receive a
payoff of P (n + 1) = (n+ 1)2 − 11 = n2 + 2n+ 1− 11 = n2 + 2n− 10
Given n donors, each citizen’s private benefit to not donating (shirking) is: S(n) = n2.
Solving for the interior Nash Equilibrium we have:
P (n+ 1) = S(n)⇔ n2 + 2n− 10 = n2 ⇒ nne = 5
As P (n+ 1) intersects S(n) from below, this is an Assurance collective action game with
two additional Nash equilibria at nne = 0 and nne = N = 100.
b. Find the socially optimal number of donors, n∗.
To find the socially optimal number of donors, we must first write down the total surplus
function for society for a given number of donors n and non-donors 100 − n:
T (n) = n(n2 − 11) + (100 − n)n2 = 100n2 − 11n
The first-order derivative of T (n) is:
∂T (n)
∂n
= 200n − 11
Which is strictly positive for all n ≥ 1 implying T (n) is always increasing in n. Hence,
n∗ = N = 100 is the socially optimal number of donors.
c. What type of Collective-Action problem is this and why?
This is an Assurance collective action problem as the P (n + 1) function intersects the
S(n) function from below. When no people are donating, S(n) is much higher than
P (n + 1) meaning it’s privately optimal for no-one to donate. When there are lots of
people donating, P (n + 1) is much larger than S(n) meaning it’s privately optimal for
everyone to donate.
Page 5 of 5
d. Suppose n is currently equal to 0 and that the government can simply convince individuals
to donate (say through other means such as tax breaks). How many people does the
government have to convince to donate in order for n = N in a Nash Equilibrium?
Recall the (unstable) interior Nash Equilibrium to this Assurance game is at nne = 5.
Thus if the government could convince n = 6 people to donate, we would reach the Nash
Equilibrium of nne = N since P (n+ 1)− S(n) grows as n grows for all n > 6.
Question 6 (20 marks)
There are four bidders interested to buy the the DVD “Lord of the Rings: Return of the King”
in an auction. Their valuations V are $5, $10, $20 and $25. Each bidder only knows her
own valuation, but all bidders know the distribution function from which all the valuations are
drawn. Assume that the bids are continuous variables (from $0 to $100).
a. Describe the optimal strategy of a bidder with valuation V in an first-price sealed-bid
auction. What bidder wins in equilibrium and at what price?
The bidder chooses his bid b to maximize his expected payoff P (b)× (V − b), where P (b)
is the probability that his bid b exceeds the largest bid r of the other bidders.
Hence, bidders face a trade-off: The higher b, the larger the probability P (b) that they
get the book, but the smaller the payoff V − b when they get the book. From the first
order condition that governs a bidders optimal bid we have:
∂P (b)
∂b
(
V − b
)
− P (b) = 0⇒ b = V − P (b)×
(∂P (b)
∂b
)
−1
< V
It directly follows that the bidder optimally shades his bid and bid b < V . Moreover, his
bid b must increase in his valuation V . As all bidders play this strategy, the bidder with
valuation V = $25 gets the book at some price b < $25.
b. Describe the optimal strategy of a bidder with valuation V in a second-price sealed-bid
auction. What bidder wins in equilibrium and at what price?
It is the bidders dominant strategy to bid b = V .
Any higher bid may lead to a negative payoff, and any lower bid may lead to the bidder
not getting the book while she could have gotten it for a price below V .
As all bidders play their dominant strategy, the bidder with valuation V = $25 gets the
book at the price b = $20.