Year 2023-2024

Game Theory and Strategic Behaviour

SET OF EXERCISES

Topic 4. SOLVING THE PRISONERS’ DILEMMA PROBLEM.

REPEATED GAMES.

KEY CONCEPTS
• The discount factor of a player.
• Repeated game with an infinite horizon.
• Grim trigger strategy in a repeated game.
• Tit-for-Tat strategy in a repeated game.

1
 Year 2023-2024

MULTIPLE CHOICE QUESTIONS

1*. Consider a repeated game where a bulb supplier must choose in every period between installing
high-quality bulbs or low-quality bulbs, while the customer must choose between paying the supplier
immediately or deferring payment. The payoffs obtained by the supplier and the customer are
displayed in the following matrix.

Customer
Pay Not pay

High quality 20, 20 -10, 25

Supplier
Low quality 25, -10 5, 5

Suppose that the game is infinitely repeated and that players discount future payoffs using a discount
factor 𝛿 ∈ (0,1). Both players use the grim trigger strategy. Could a cooperative equilibrium where
the supplier installs high-quality bulbs and the customer pays immediately be sustained?

a) No, because both players have a dominant strategy: the supplier always prefers to install
low-quality bulbs and the customer always prefers to defer payment.
b) Yes, a cooperative equilibrium can be sustained as long as 𝛿 > 1/4.
c) Yes, a cooperative equilibrium can be sustained as long as 𝛿 = 1/5 .
d) Yes, a cooperative equilibrium can be sustained for any value of 𝛿.

2. In the electricity market, two companies are deciding their pricing policy. They can set high or low
prices. The following matrix presents the profits of each of them according to the chosen actions.

LUMEN
High price Low price

High price 22, 22 18, 25

IBER
Low price 25, 18 20, 20

Suppose that the game is infinitely repeated and that players discount future payoffs using a discount
factor 𝛿 = 0.7. Both players use the Tit-for-Tat strategy. What is the present discounted value of
cooperating (i.e., to obtain the efficient result in every period) and what is/are the present discounted
value/s of the deviation/s?

2
 Year 2023-2024

a) Present value of cooperation: 79.12, present value of deviations: 67.34 and 63.72.
b) Present value of cooperation 80.15, present value of deviation: 67.34.
c) Present value of cooperation 73.33, present value of deviations: 71.66 and 73.72.
d) Present value of cooperation 92.23, present value of deviations: 75.3 and 67.34.

3
 Year 2023-2024

SHORT QUESTIONS

3. Explain why repeated games with an infinite horizon are equivalent to repeated games with an
undefined horizon where there is a constant probability that the game is repeated one more period.

4. “Cooperation is the unique Nash equilibrium outcome in the repeated prisoners’ dilemma with
infinite horizon.” True or false? Explain your answer.

5*. “The grim trigger combination of strategies in a Prisoners’ Dilemma is not a Nash equilibrium in
a repeated game with infinite horizon if any player has a discount rate of 𝛿 = 0.” True or false?
Explain your answer.

4
 Year 2023-2024

PRACTISING THE BASIC CONCEPTS

6. (Tax setting). Game theory is a good tool for analysing what has happened in recent years with the
tax policy in the European Union, in particular with regard to corporation tax. Suppose two countries,
A and B, can impose either a high or a low tax, and they announce it simultaneously. The following
payoff matrix shows the tax revenue of both countries.

Country B
High tax Low tax
High tax 6, 6 1, 7
Country A
Low tax 7, 1 3, 3

a) What tax would each country set if they had to decide only once, and this strategic situation
would not be repeated?
b) If the tax rate can be changed every year, what strategies could materialize the cooperative
solution (and avoid tax reductions)?

7*. (Why do we only consider one deviation with the grim trigger strategy?) Consider the following
payoff matrix of a simultaneous game with complete information.

C NC
C 𝑐, 𝑐 𝑏, 𝑎
NC 𝑎, 𝑏 𝑛, 𝑛

a) Obtain the relation among 𝑎, 𝑏, 𝑐 and 𝑛 for this game to have a Prisoners’ Dilemma structure.
Obtain the Nash equilibrium if this game is only played once.
b) Now consider that the game is repeated with an infinite horizon and that players consider the
grim trigger strategy to reach the efficient outcome of the game as an equilibrium. Assume
that both players have the same discount factor 𝛿.
i. Calculate the present value (payoff stream) of Player 1 if they both follow the
strategy.
ii. Calculate the present value (payoff stream) of Player 1 if he/she deviates from
the strategy in the first period, 𝑡 = 1.
iii. For which range of values of 𝛿 is this deviation profitable?

5
 Year 2023-2024

c) Consider now that Player 1 deviates in period 𝑡 = 4 instead.

i. Calculate the present value (payoff stream) of Player 1 if he deviates from the
grim trigger strategy in the fourth period.
ii. For which range of values of 𝛿 is this deviation profitable?
iii. Has the threshold on 𝛿 changed between section b) and section c)?

6
 Year 2023-2024

STANDARD EXERCISES

8. (Tacit collusion in the optical fibre market. Price competition-Bertrand). In the optical fibre market,
two firms are deciding their pricing policy. They must simultaneously decide whether to set a high
price or a low price. If both firms decide to set a high price, their profits will be of 10 million euros for
each firm. If both firms decide to set a low price, their profits will be of 5 million euros for each firm. If
a firm sets a high price and the other firm sets a low price, profits will be of 0 euros for the firm that
sets the high price and 15 million euros for the firm that sets the low price.

a) Consider that this game is only played once. Represent the game in matrixial form. Calculate
the Nash equilibrium and the equilibrium profits for each firm. Discuss the efficiency of this
result.
b) Describe a self-binding agreement that allows firms to obtain a more profitable equilibrium.
Calculate the profits of such agreement. Describe the strategy and calculate the critical
discount factor for both firms setting a high price to be a Nash equilibrium of the repeated
game. Explain your answer.
c) Consider that a firm can violate the agreement and set a low price to obtain greater profits
for 3 periods before being discovered. Is it possible to achieve the cartel (cooperative)
equilibrium through the self-binding agreement if both players’ discount factor is 𝛿 = 0.2?
What profits would each firm obtain in equilibrium?
d) Discuss to what extent you agree or disagree with the following statement: “If both players’
discount factor is of 𝛿 = 0.9, the unique Nash equilibrium of the repeated game is that both
firms set a high price.”

(Exam GECO, June 2023)

9. (Tacit collusion between counties exporting a mineral. Competition in quantities-Cournot). You are
one of the OREC member nations: the organization of rhodium exporting countries designed to
artificially restrict the quantity of product in an attempt to generate profits. The demand conditions are
such that the different combinations of production and annual profits are shown in the following matrix:

Russia
Produce 1000 Tm Produce 2000 Tm

Produce 1000 Tm 500, 500 250, 700

South Africa
Produce 2000 Tm 700, 250 400, 400

a) What is the equilibrium result if this game is only played once? Is this the best result for each
country or could they achieve something better? Explain your answer.
b) If South Africa and Russia were able to collude successfully, how many tons would each
country produce and what profits would each one make in one year?

7
 Year 2023-2024

c) Suppose that both countries agree to produce 1000 tons while the other produces 1000 tons
as well. But in the event that one of them produces 2000 tons, the Tit-for-Tat strategy will be
applied. What flow of profits will be generated for each country if both countries adopt this
punishment strategy? Express your answer as a sum.
d) Consider that the discount factor for both countries is 𝛿 = 0.7. What would they do? Follow or
break the agreement?

10. (The contribution to a public good can be an equilibrium). Two individuals must simultaneously
decide in every period their contribution to a public good in a repeated situation with infinite horizon.
Individual contributions 𝑥! , can only adopt two values: 0 or 10, where 𝑖 = 1,2. The payoff function of
each player, given a pair of contributions, is: 𝑢! = 10 – 𝑥! + 𝒂(𝑥" + 𝑥# ), where 𝒂 = 0.6 is the
marginal individual return of the public good. Both players have the same discount factor d.
a) Draw the payoff matrix of the stage-game. Is it possible to obtain the efficient outcome in every
period as a Nash equilibrium of the repeated game in which both players use Tit-for-Tat
strategies?
b) What happens if the return of the public good, a, increases? Is it easier or more difficult to
sustain cooperation? Provide some intuition about the results you obtain. Assume that 1 >
𝒂 > 0.5.

In press, a Valencian nobel? https://www.levante-emv.com/economia/2009/10/14/nobel-sabor-

valenciano-13201567.html

Tribunal de las Aguas: https://tribunaldelasaguas.org/es/

11. (Pharmaceutical pricing and fines on collusive practices). The profits made by two firms belonging
to the pharmaceutical sector depend on the prices that both firms set, and such profits are displayed
in the following payoff matrix:

Firm B
High price Low price

High price 60, 60 30, 80

Firm A
Low price 80, 30 40, 40

a) Let us assume that this game is repeated indefinitely and that both firms sign a secret
agreement to set high prices (price collusion) and obtain higher profits. Which punishment
strategies could they use between them and how would they work? Using a punishment

8
 Year 2023-2024

strategy, determine for which discount factor 𝛿 could this agreement be sustained. Explain
your answer.
b) Consider that the Competition Authority discovers this agreement in the first period and sets
each firm a fine of 150 in that period. Suppose that both firms have a discount factor of d =
0.6. Can such fine deter firms from continuing setting high prices? What would the minimum
dissuasive fine have to be?

In press, about the 10 greatest fines on cartels: https://www.eleconomista.es/empresas-

finanzas/noticias/7050410/10/15/Las-10-mayores-multas-de-Competencia-de-la-historia.html

12. (Asymmetric teamwork with Tit-for-Tat strategies). In a firm with two partners, each one
independently decides how much effort to exert in their work. The level of output, which is equally
shared between the two partners, is given by 𝑂 = 6(𝑒" + 𝑒# ), where 𝑒" is the effort chosen by
player 1 and 𝑒# is the effort chosen by player 2. The individual cost of effort of player 1 is given by
the function 𝑐(𝑒" ) = 4𝑒" and for player 2 is given by the function 𝑐(𝑒# ) = 4𝑒# , if player 2 exerts an
effort greater or equal than his opponent and by 𝑐(𝑒# ) = 4𝑒# – 2, if he exerts a smaller effort than
his opponent. Suppose that efforts can take two values: 1 or 2.

a) Draw the payoff matrix of the stage game. Define the Tit-for-Tat strategy for the repeated
game. Is it possible to obtain the efficient outcome in every period as a Nash equilibrium of
the repeated game in which both players use this strategy?
b) Assume the same situation as in the previous section, except that now player 2 is inequity
averse with a utility function 𝑈# (𝑥" , 𝑥# ) = 𝑥# − 0.4 𝑚𝑎𝑥{𝑥# – 𝑥" , 0}, where 𝑥" and 𝑥# are the
material payoffs of both players. Under which conditions is it possible to obtain the efficient
equilibrium during every period using the Tit-for-Tat strategy? Explain your answer.

13*. (Exam exercise). Consider a trade war between USA and China. They must decide, without
knowing what will the other country do, if they impose tariffs or not impose tariffs. If both countries
impose tariffs, each one will receive a payoff of 2. If no country imposes tariffs, each one will receive
a payoff of 5. If only one country imposes tariffs, the payoff for that country will be of 10 and the payoff
for the country that has not imposed tariffs will be of -1.

a) What type of game is the strategic interaction described by the payoff matrix? Describe the
characteristics of this type of games.
b) Define what is a Pareto-efficient outcome. Identify all outcomes in this game that satisfy this
definition.
c) Consider that this game is repeated indefinitely and that the discount factor for both players is
𝛿 = 0.7. Each country follows the Tit-for-Tat strategy. Describe this strategy. Would it be
possible to achieve the free trade outcome (without tariffs) as a Nash equilibrium of the
repeated game? Explain your answer.

9
 Year 2023-2024

d) Explain to what extent you agree with the following statement: “If each country followed the
strategy of not imposing tariffs in each period, independently of what happened before, the
free trade outcome would be a Nash equilibrium of the repeated game”. Explain your answer.

(Exam GECO, January 2023)

10
 Year 2023-2024

ADVANCED EXERCISES

14*. (Tax compliance) In this exercise we are going to assume that the economic situation is such
that each player knows that the probability with which the game continues next period is 𝑝.

Each year the following strategic situation between a contributor and the Tax Agency (both with 𝛿 =
1) happens. The contributor can pay (P) or not pay (NP). The Tax Agency chooses between inspecting
(I) or not inspecting (NI). The game is represented by the following payoff matrix:

Tax Agency
NI I

P -2, 2 -3, 1
Contributor
NP 0, 0 -4, 3

a) Consider that the contributor is young and healthy such that the probability of continuing
working (and therefore interacting with the Tax Agency) is relatively high, 𝑝 = 0.8. Explain
which strategy could the players follow for (P, NI) with payoffs (-2, 2) to be a Nash equilibrium
of the repeated game.
b) Now consider that the contributor is an old man with a history of cardiovascular events and,
therefore, 𝑝 = 0.2. Could the same equilibrium of the repeated game be reached?

15. (Repeated Cournot game). Consider a Cournot duopoly where two firms compete producing
identical goods. Consider that in the market in which these firms compete the unit price is 𝑝 = 10 −
𝑋, where 𝑋 is the total market production (𝑋 = 𝑥" + 𝑥# ) and 𝑥" and 𝑥# are the units produced by each
firm 1 and 2 respectively. Firms compete choosing their production simultaneously. Suppose, for
simplicity, that marginal production costs are 0, such that firm 𝑖’s profit function 𝜋! is given by firm 𝑖’s
revenue. Additionally suppose that both firms have the same discount factor 𝛿.
a) Calculate the best-response functions (reaction functions) and the Nash equilibrium if this
game is only played once (price, quantities and profits).
b) What would the result be if there was only one firm in this market (monopoly)? (Price,
quantities and profits).
c) If the game were repeated with an infinite horizon, what would the collusive agreement in
terms of quantities between the two firms be (strategies)? How much would firms earn with
such an agreement?
d) If a firm thinks about breaking the agreement, how would it deviate? What would the profits of
the deviation be?
e) How could the “betrayed” firm punish the other firm?

11
 Year 2023-2024

STRATEGIC SITUATION – GAME IN THE NEWS

Read the following news article (EP, 2016) and answer the following questions:

Brussels imposes fines of 3 billion euros on five truck manufacturers

MAN, Volvo/Renault, Daimler, Iveco and DAF have admitted their involvement in the 14-year cartel.

On Tuesday, the European Commission announced a fine of 2.93 billion euros on five truck
manufacturers - MAN, Volvo/Renault, Daimler, Iveco and DAF - which it accuses of setting the selling
prices of their vehicles and passing on to the buyer the costs of complying with emissions standards.

This is the largest fine imposed by Brussels on a cartel up to the date, after the 1.47 billion fine
imposed in 2012 against an agreement among seven television manufacturers in order to lower
prices.

The five sanctioned manufacturers have acknowledged their involvement in the cartel, which lasted
fourteen years, between 1997 and 2011, so the EU executive has agreed to reduce the fines by 10%
and has even completely forgiven the fine imposed on MAN, as the company that confessed to the
existence of the cartel.

"This is a record fine for a serious infringement", warned the Commissioner for Competition,
Margrethe Vestager, who reminded in a press conference that more than 30 million lorries circulate
on the roads of the European Union.

Vestager has insisted that “it is not acceptable” that a group of companies that together manufacture
nine out of ten medium and heavy trucks in Europe, agree on sales prices “instead of competing with
each other”'.

Specifically, DAF will have to pay 752,679,000 euros, Volvo (owned by Renault) will have to pay a
fine of 670,488,000 euros, Iveco a fine of 494,606,000 euros and Daimler, which is imposed the
largest amount, 1,008,766,000 euros.

The Commissioner has indicated that there is a sixth manufacturer - Scania - whose behaviour is still
being investigated, so it has not been included in Tuesday's decision and the pronouncement of
Brussels is pending.

Scania decided not to participate in this “friendly” procedure, which has ended with reduced fines for
manufacturers who have admitted wrongdoing, the Commissioner explained. Vestager added that
the procedure continues for Scania, for which the deadline to present its defence is now reopened.

The most serious irregularities detected by Brussels in its investigation concern the agreement
between the cartel participants to fix a “'raw price scale” for medium (between 6 and 16 tonnes) and
heavy (over 16 tonnes) vehicles. These are the ex-factory prices which then condition prices in the
sector.

12
 Year 2023-2024

However, Brussels has not provided an estimation of the average impact of this agreement on the
final price of trucks, as this is not determined solely on the basis of the ex-factory scale, but is affected
by other factors.

They also agreed on a timetable for delaying the introduction of the necessary technology in their
vehicles to meet EU emissions standards.

The third factor taken into account by the Commission in setting the fine is that the five major brands
passed on to customers the costs of introducing emission control systems in their products.

Reference: EP (2016, 19th July). Brussels imposes 3 billion fine on five truck manufacturers. Cinco
Días, El País [en línea].
https://cincodias.elpais.com/cincodias/2016/07/19/empresas/1468926817_807600.html [2016, 19th
July]

Questions
a) Describe the strategic situation as if there were only two players. Is it a simultaneous game or
a sequential game?
b) Present the stage game in strategic form and propose payoffs for it to have the structure of a
Prisoners’ Dilemma. Calculate the Nash equilibrium. Discuss its efficiency.
c) Consider now that the game is repeated with an infinite horizon. Describe a conditional
punishment strategy that could help players sustain an efficient outcome as a Nash equilibrium
of the repeated game. Under which conditions would this work? Discuss the efficiency of the
NE of the repeated game if these conditions are satisfied.

13