BE

510 Business Economics 1 - Autumn 2021

Problem Set 2

1. (a) Find the Nash equilibria of the Player 2

game shown on the right.
Left Centre Right
(b) Repeat (a) when the payoffs in Up 350, 350 350, 250 1000, 0
the (Down, Right) cell (500,
Player 1 Middle 250, 350 550, 550 0, 0
500) are replaced by (600,
Down 0, 1000 0, 0 500, 500
600).
(c) The games in (a) and (b) have been studied in decision-making experiments involving
subject volunteers and actual monetary incentives. What would you predict participants
typically choose in these two games and why?

2. Consider a variant of the ‘battle of the sexes’ game. Bob

Each player has a preferred location but the more Ballet Football
important consideration for both is that they Ballet 3 , 2 1 , 1
show up at the same location. The possible payoffs Alice
Football 0 , 0 2 , 3
for each player are 0, 1, 2 or 3.
Now construct another variant with the following characteristics: (i) Bob’s payoffs from the
different outcomes remain as before. (ii) Alice still prefers the ballet to the football but the
more important consideration for her now is that she avoids Bob. (Reuse the payoffs 0, 1, 2
and 3 for Alice but put them in the cells appropriate for the new situation.) Does this new
game have a Nash equilibrium?

3. (a) Find the Nash equilibria and the Pareto-efficient outcomes in these two games.
Note: An outcome is Pareto-efficient if it is impossible, by switching to another outcome, to make one
player better off without making the other player worse off.
[A] Player 2 [B] Player 2
Left Right Left Right
Up 3 , 4 1 , 3 Up 2 , 2 4 , 1
Player 1 Middle 0 , 2 8 , 8 Player 1 Middle 3 , 1 1 , 0
Down 2 , 0 9 , 1 Down 1 , 3 6 , 1

(b) Now consider a game in which, as before, player 1 chooses the row (Up, Middle or Down)
and player 2 chooses the column (Left or Right), but where there is also a third player
who chooses the table (A or B). Player 3’s payoffs are shown below as the third number
in each cell. The payoffs for player 1 and player 2 remain unchanged. Find the Pareto-
efficient outcomes and the Nash equilibria when all the players move simultaneously.
 BE 510 Business Economics 1 - Autumn 2021

[A] Player 2 [B] Player 2

Left Right Left Right
Up 3 , 4 , 6 1 , 3 , 3 Up 2 , 2 , 4 4 , 1 ,5
Player 1 Middle 0 , 2 , 2 8 , 8 , 8 Player 1 Middle 3 , 1 , 4 1 , 0 , 4
Down 2 , 0 , 0 9 , 1 , 2 Down 1 , 3 , 3 6 , 1 , 0

4. Four friends share a flat. After breakfast there is no more milk in the fridge. All leave the house
to pursue different activities during the day, and all will be very busy until the evening. But
someone ought to buy new milk (although it’s a hassle). The payoffs for each of the four
players are:

Flat mates
At least one of the None of the others
others buys milk buys milk
Go and buy some milk −1 −1
Me
Don’t buy milk 2 −3

(a) Determine the Nash equilibria and the Pareto-efficient outcomes of this game.
(b) Write down the normal form payoff matrix of the two-player version of this game.

Player 2
Left Centre Right
Up 350, 350 350, 250 1000, 0
Player 1 Middle 250, 350 550, 550 0, 0
Down 0, 1000 0, 0 500, 500
 BE 510 Business Economics 1 - Autumn 2021

Problem Set 2 - Solutions

No guarantees for correctness. If you find errors in the proposed solutions, please let us know.

1. (a) Find the Nash equilibria of the following game.

Player 2
Left Centre Right
Up 350, 350 350, 250 1000, 0
Player 1 Middle 250, 350 550, 550 0, 0
Down 0, 1000 0, 0 500, 500

There are two equilibria: (Up, Left) and (Middle, Centre).

(b) Repeat (a) when the payoffs in the (Down, Right) cell (500, 500) are replaced by (600,
600).
This game is of course almost identical to the first one. The change in the (Down, Right)
cell does not affect the equilibrium analysis and so we get the same results as before.
(c) The games in (a) and (b) have been studied in decision-making experiments involving
subject volunteers and actual monetary incentives. What would you predict participants
typically choose in these two games and why?
The experiments are reported in Cooper, R.W., D.V. DeJong, R. Forsythe and T.W. Ross
(1990): Selection criteria in coordination games: Some experimental results. American
Economic Review, 80(1), 218–233. Using these and some other similar games, the
researchers first of all found that in the vast majority of cases subjects chose actions that
are consistent with Nash equilibrium play. However, the choices in games (a) and (b)
were very different. Focusing on the results from later rounds (after subjects had gained
some experience), in game (a) the typical outcome was the (Middle, Centre) equilibrium
and in game (b) the typical outcome was the (Up, Left) equilibrium. Thus, a rather small
change in the payoffs of the dominated(!) strategy has a very strong impact on which
equilibrium players select.
Given the presence of two Nash equilibria both games are characterized by a certain
coordination problem. However, since the (Middle, Centre) equilibrium payoff-
dominates the (Up, Left) equilibrium many theorists would argue that players ought to
be able to resolve this coordination problem by focusing on the payoff-dominant (Middle,
Centre) equilibrium, in either game. Of course, players might be worried about the
possible zero payoff that one might incur when one chooses Middle or Centre—in some

1
Problem Set 2 - Solutions BE 510 Business Economics 1 - Autumn 2021

sense Up or Left is the ‘safer’ option. The zero payoff, however, only becomes relevant if
the other player picks the strictly dominated action (Down or Right). Therefore, when we
observe that players select Up or Left the reason is, presumably, that such players attach
at least some positive probability to the event that the other party chooses the dominated
action.
Now, the experimental evidence indicates that people are much more worried about the
other player choosing the dominated action in game (b) than in game (a), even though the
payoffs are almost identical. A plausible explanation for this is that in game (a) the
(Middle, Centre) equilibrium yields a Pareto efficient outcome, whereas it doesn’t in game
(b) because in game (b) the (Middle, Centre) outcome is Pareto dominated by the (Down,
Right) outcome. Thus, subjects seem to think that others might try to ‘cooperate’ by
playing Down or Right in game (b). In game (a) there is no scope for such ‘cooperation’
and subjects tend not to worry about the other player choosing the dominated action.
Standard game theoretical analysis assumes common knowledge of rationality and under
this assumption strictly dominated actions are completely ruled out, whatever the precise
values of the dominated payoffs are in a particular game. One could therefore say that the
experiment demonstrates the failure of the common knowledge of rationality assumption.
More precisely, the experiment shows something more subtle: Whether or not players
expect others to dismiss dominated strategies can depend on the precise nature of the
payoffs from those strategies relative to other payoffs in the game. Players do not appear
to think that others are completely irrational and ignore obvious dominance but they
apparently do think that others may aim for a Pareto efficient outcome when they see
one. Interestingly, these beliefs about others do not necessarily match the actual behavior
of others: The observed frequency of Down/Right is close to zero even in the (b) game.

2. Consider a variant of the ‘battle of the sexes’ game. Bob

Each player has a preferred location but the more Ballet Football
important consideration for both is that they Ballet 3 , 2 1 , 1
show up at the same location. The possible payoffs Alice
Football 0 , 0 2 , 3
for each player are 0, 1, 2 or 3.
Now construct another variant with the following characteristics: (i) Bob’s payoffs from the
different outcomes remain as before. (ii) Alice still prefers the ballet to the football but the
more important consideration for her now is that she avoids Bob. (Reuse the payoffs 0, 1, 2
and 3 for Alice but put them in the cells appropriate for the new situation.) Does this new
game have a Nash equilibrium?

2

Problem Set 2 - Solutions BE 510 Business Economics 1 - Autumn 2021

The new game is

Bob
Ballet Football
Ballet 1 , 2 3 , 1
Alice
Football 2 , 0 0 , 3

This game has no Nash equilibrium. This must be so because Alice’s and Bob’s interests are
completely at odds with each other: While Bob best responds by choosing the same action as
Alice, Alice’s best response is to always take the other action. Hence, no combination of actions
can lead to mutual best responses.

3. (a) Find the Nash equilibria and the Pareto-efficient outcomes in these two games.
Note: An outcome is Pareto-efficient if it is impossible, by switching to another outcome, to make one
player better off without making the other player worse off.
[A] Player 2 [B] Player 2
Left Right Left Right
Up 3 , 4 1 , 3 Up 2 , 2 4 , 1
Player 1 Middle 0 , 2 8 , 8 Player 1 Middle 3 , 1 1 , 0
Down 2 , 0 9 , 1 Down 1 , 3 6 , 1

Nash equilibria: Pareto-efficient outcomes:
§ Game A: {Up, Left}, {Down, Right}. § Game A: (8, 8) and (9, 1).
§ Game B: {Middle, Left} § Game B: (2, 2), (1, 3) and (6, 1).

(b) Now consider a game in which, as before, player 1 chooses the row (Up, Middle or Down)
and player 2 chooses the column (Left or Right), but where there is also a third player
who chooses the table (A or B). Player 3’s payoffs are shown below as the third number
in each cell. The payoffs for player 1 and player 2 remain unchanged. Find the Pareto-
efficient outcomes and the Nash equilibria when all three players move simultaneously.

[A] Player 2 [B] Player 2

Left Right Left Right
Up 3 , 4 , 6 1 , 3 , 3 Up 2 , 2 , 4 4 , 1 ,5
Player 1 Middle 0 , 2 , 2 8 , 8 , 8 Player 1 Middle 3 , 1 , 4 1 , 0 , 4
Down 2 , 0 , 0 9 , 1 , 2 Down 1 , 3 , 3 6 , 1 , 0

Nash equilibria: {Up, Left, A}, {Down, Right, A} and {Middle, Left, B}.
Pareto-efficient outcomes: (8, 8, 8) and (9, 1, 2).

3

Problem Set 2 - Solutions BE 510 Business Economics 1 - Autumn 2021

4. Four friends share a flat. After breakfast there is no more milk in the fridge. All leave the house
to pursue different activities during the day, and all will be very busy until the evening. But
someone ought to buy new milk (although it’s a hassle). The payoffs for each of the four
players are:

Flat mates
At least one of the None of the others buys
others buys milk milk
Go and buy some milk −1 −1
Me
Don’t buy milk 2 −3

(a) Determine the Nash equilibria and the Pareto-efficient outcomes of this game.
The above game has four Nash equilibria. In each of these, exactly one of the four friends
buys the milk. The equilibria differ simply in who is doing it. The Nash equilibria are all
Pareto-efficient. When more than one person buys milk, it would be possible to obtain a
Pareto improvement by reducing the number of milk buyers to 1: The remaining person
buying milk would not be worse off and the former milk buyers would now earn 2 instead
of −1. When nobody buys milk, it is clearly better for everyone if somebody does it.
Some additional remarks: This type of game is known as the volunteer’s dilemma. It reflects a setting
in which one person is enough to resolve a particular problem. The difficulty is that taking the required
action is costly and thus the incentive is not to volunteer unless there really is nobody else who is
volunteering instead.
(b) Write down the normal form payoff matrix of the two-player version of this game.
Player 2
Buy milk Don’t buy milk
Buy milk −1 , −1 −1 , 2
Player 1
Don’t buy milk 2 , −1 −3 , −3

Some additional remarks: The two Nash equilibria are (Don’t buy milk, Buy milk) and (Buy milk, Don’t
buy milk). This is an example for the so-called chicken game. This game is often associated with a very
different story based on the famous James Dean movie “Rebel Without a Cause”. In one scene James Dean
and his rival drive their (stolen) cars at high speed towards a cliff. The rule is that whoever loses his nerves
first and hits the brakes is a ‘chicken’. If both hit the brakes at the same time, it’s a draw. Going over the
cliff is not particularly advisable.
Red car
Another story is about two aggressive
Swerve Drive on
car drivers about to enter a short,
narrow part of the road (or a short, Swerve A draw Victory for Red,
Blue car

narrow bridge) from opposite ends at Blue loses

the same time. Whoever swerves and lets
Drive on Victory for Blue, Collision and
the other car pass is the loser.
Red loses bad injury