BE

510 Business Economics 1 - Autumn 2021

Problem Set 4

1. A small dairy farmers’ cooperative negotiates a contract to supply milk to DeliYumm, a large
chain of quality food stores. The cooperative requires a minimum price of 24c, i.e. rather than
accept a price below 24c it would refuse a deal with DeliYumm and implement an alternative
business plan. After some negotiations the two parties settle on a price of 35c. Following the
agreement the cooperative makes sizeable investments in order to change aspects of its
delivery system so that it complies with DeliYumm’s unusual requirements.
A couple of years later another meeting is scheduled to negotiate a renewal of the contract.
The situation is unchanged (the cooperative can effectively obtain a price of 24c elsewhere)
although the cooperative feels it needs at least 30c to recover the mentioned investments in
a reasonable time.
DeliYumm’s negotiating team prepares a written offer for the meeting. In doing so the team
members have to take into account and balance several issues:
− They must get the offer pre-approved by their superiors. Any price up to, but no price
above 35c will get such an approval.
− Any later change to the pre-approved price as a result of the meeting will be refused by the
superiors.
− The cooperative’s representatives are obviously free to reject an unsatisfactory offer.
− If the team’s offer is rejected the superiors will not be happy: They will evaluate the team’s
performance as if a price of 36c had been agreed.
− If the team’s offer is accepted the superiors will evaluate the outcome in terms of the agreed
price: The lower the better.
(a) Represent the game between the DeliYumm team and the cooperative in extensive form
and in normal form. Assume the following:
− The cooperative seeks to maximize profits. The members of DeliYumm’s negotiation
team seek to maximize their performance as judged by their superiors.
− For simplicity: The offer can only be 25c, 30c or 35c. Furthermore, if a deal is reached
the price level has no ramifications for the future beyond the duration of the contract.
− All information provided is common knowledge.

(b) Find all pure-strategy Nash equilibria and subgame-perfect equilibria.

 BE 510 Business Economics 1 - Autumn 2021

2. Consider the following game in normal form.

(a) Find all pure-strategy Nash equilibria.
Player 2
(b) Suppose the game is repeated infinitely often. Left Centre Right
Determine the minimum level of the discount Up 5, 5 0, 0 −1, 0

Player 1
factor required to sustain the Middle-Centre
Middle 0, 0 10, 10 −3, 0
outcome under a suitable trigger strategy.
Down 0, −1 12, −2 −1, 0
(c) Suppose the game is played twice. Ignoring any
discounting of future payoffs, find a Nash equilibrium such that Middle-Centre is played
in the first round. Explain.


Problem Set 4 - Solutions BE 510 Business Economics 1 - Autumn 2021

Problem Set 4 - Solutions

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

1. A small dairy farmers’ cooperative negotiates a contract to supply milk to DeliYumm, a large
chain of quality food stores…
(a) Represent the game between the DeliYumm team and the cooperative in extensive form
and in normal form…
This is a game where we are not given numerical payoffs. However, we do know the range
of possible outcomes and their relative attractiveness for each party. The worst outcome
for the negotiating team would be a rejection. Let’s denote the team’s payoff associated
with this outcome as 𝑎! . The other possibilities are that a deal is struck at a price of 35c
(𝑎" ), 30c (𝑎# ) or 25c (𝑎$ ). We know that 𝑎$ > 𝑎# > 𝑎" > 𝑎! .
The best outcome for the cooperative would be a deal at a price of 35c (𝑏$ ), followed by
30c (𝑏# ), 25c (𝑏" ) and “no deal” (𝑏! ), that is 𝑏$ > 𝑏# > 𝑏" > 𝑏! . Note here that although
the cooperative would not be very happy about a 25c deal the text suggests that it would
not be best advised to reject such an offer. The reason is that the investments it seeks to
recover are clearly specific to the arrangements with DeliYumm. If the contract was not
renewed, it had to write off the investment expenditures or, even worse, incur additional
costs to undo the changes in its delivery system. This fact weakens its bargaining position
here.
Additional remark: This is an effect known as the holdup problem.
To simplify things, we are free to normalize the range of payoffs in some way. For example,
we could define that the worst outcome for each party is associated with a payoff of 0 and
that the best outcome for each party is associated with a payoff of 1. Since preferences
can be represented by different utility functions this normalization is not an additional
assumption (but going on further and arbitrarily assuming a value of, say, 0.8 for the
second-best outcome would be a problem from the perspective of expected utility theory,
which is underlying standard game theory).
With this in mind, we get the extensive-form representation of the game that is shown on
the next page.
The DeliYumm team offers 25, 30, or 35, and the cooperative then subsequently decides
whether or not to accept the given offer. A rejection yields a (normalized) payoff of zero
for both parties. When the offer is accepted positive payoffs are obtained, and DeliYumm
prefers lower offers while the cooperative prefers higher offers.

1

Problem Set 4 - Solutions BE 510 Business Economics 1 - Autumn 2021

1, 𝑏!
accept
Cooperative
reject 0, 0

25
accept 𝑎" , 𝑏"
DeliYumm
Cooperative
Team
30 reject 0, 0

35

accept 𝑎! , 1
Cooperative
reject 0, 0

The normal form representation of this game involves three pure strategies for DeliYumm
(25, 30 and 35) and eight pure strategies for the cooperative: For any offer the
cooperative could either accept or reject the offer, and therefore a complete plan of action
considers all possible combinations of accept/reject responses. Thus, a strategy specifies
the response to an offer of 25, the response to an offer of 30 and the response to an offer
of 35, for example: “25: Accept, 30: Reject, 35: Reject” (shorthand: ARR) or “25: Reject,
30: Accept, 35: Reject” (shorthand: RAR). Hence, the normal form looks as follows.

DeliYumm
Offer 25 Offer 30 Offer 35
AAA 𝑏! , 1 𝑏" , 𝑎" 1, 𝑎!
AAR 𝑏! , 1 𝑏" , 𝑎" 0, 0
ARA 𝑏! , 1 0, 0 1, 𝑎!
Cooperative

ARR 𝑏! , 1 0, 0 0, 0
RAA 0, 0 𝑏" , 𝑎" 1, 𝑎!
RAR 0, 0 𝑏" , 𝑎" 0, 0
RRA 0, 0 0, 0 1, 𝑎!
RRR 0, 0 0, 0 0, 0

(b) Find all pure-strategy Nash equilibria and subgame-perfect equilibria.

Our usual analysis produces 7 pure-strategy Nash equilibria: AAA, AAR, ARA and ARR are
all best responses to an offer of 25 and vice versa; RAA and RAR are best responses to an
offer of 30 and vice versa; and RRA is a best response to an offer of 35 and vice versa.

2

Problem Set 4 - Solutions BE 510 Business Economics 1 - Autumn 2021

Subgame-perfect Nash equilibria: The cooperative’s best response in stage 2 of the game
is always ‘Accept’. Anticipating this, DeliYumm makes the lowest offer, 25. Thus, the
strategy pair (25, AAA) is the unique subgame-perfect equilibrium.
Some additional remarks: This game is an example of the so-called ultimatum game. It is one of the
simplest models of a bargaining situation between two parties. The idea is that player 1 makes a take-it-
or-leave-it offer to player 2, and if player 2 does not accept player 1’s offer, the negotiations are
terminated. This describes a situation where one party is very dominant: If you find yourself in a position
where you can make a take-it-or-leave-it offer, this implies that you have a lot of bargaining power.
Numerous decision-making experiments have been conducted on this and closely related games. The
typical outcome deviates from the subgame-perfect prediction and among economists this has instigated
a whole new research agenda on how people’s decisions are affected by fairness concerns.

2. Consider the following game in normal form.

Player 2
Left Centre Right
Up 5, 5 0, 0 −1, 0
Player 1 Middle 0, 0 10, 10 −3, 0
Down 0, −1 12, −2 −1, 0

(a) Find all pure-strategy Nash equilibria.

There are two pure-strategy equilibria: (Up, Left) and (Down, Right).
(b) Suppose the game is repeated infinitely often. Determine the minimum level of the
discount factor required to sustain the Middle-Centre outcome under a suitable trigger
strategy.
The idea of a trigger strategy is that each player repeatedly chooses a ‘cooperative’ action
unless the other player deviates from the cooperative plan. If a deviation occurs, the
punishing player reverts, forever, to an action consistent with a Nash equilibrium of the
stage game. Anticipating this, the best response of the punished player from that point
onwards is to go along with that stage game Nash equilibrium (otherwise it wouldn’t be
a Nash equilibrium).
The obvious cooperative combination here is (Middle, Centre). However, what about the
stage game Nash equilibrium? Recall that we found two pure-strategy equilibria. In one
of these (the “good” equilibrium) both players earn 5 while in the other one (the “bad”
equilibrium) they earn −1 and 0. Which one should the players use for their trigger
strategies? Well, since the idea of a trigger strategy is that players stick to the cooperative
outcome because of the looming threat of the low-payoff stage-game equilibrium, the
most effective way to sustain cooperation is to make the threat as severe as possible.
Therefore, the bad equilibrium (Down, Right) should be used.
Note that in this game the only player who needs to be kept in check is player 1 because
for player 2 Centre actually is a best response to player 1’s Middle. Player 1, on the other
hand, can get 12 instead of 10 by going Down. In the repeated game player 1’s present
3

Problem Set 4 - Solutions BE 510 Business Economics 1 - Autumn 2021

value payoff from Middle is 10⁄(1 − 𝛿), given that player 2 plays Centre, and player 1’s
present value from (Down, Centre) once and (Down, Right) thereafter is 12 − 𝛿 ⁄(1 − 𝛿).
The former is greater than the latter if
10 𝛿 2
> 12 − or 𝛿 > = 0.1538.
1−𝛿 1−𝛿 13

(c) Suppose the game is played twice. Ignoring any discounting of future payoffs, find a Nash
equilibrium such that Middle-Centre is played in the first round. Explain.
We make use of the fact that the two Nash equilibria of the one-shot version of the game
(see part (a) above) are very different: (Up, Left) is quite nice, whereas (Down, Right) is
an outcome the players would rather avoid. Therefore, how about using the bad Nash
equilibrium as a “threat” and the good one as a “reward”? The idea is: If both players
behave in a desirable way in the first round, they then reward each other by playing the
good equilibrium in the second round—but if one of them does (or both do) something
undesirable, then they go for the bad equilibrium and expect the other player to do the
same. Importantly, the good and the bad round-2 outcomes are not just any good or bad
outcomes, but they are Nash equilibria, and this means that the reward or the threat are
credible!
Thus, each player’s strategy for round 2 could be “Play Up[Left] if the outcome of round
1 has been (10, 10); otherwise play Down[Right].” Given these strategies for round 2, the
anticipated overall payoffs at the beginning of round 1 look as follows.
Player 2
Left Centre Right
Up 4, 5 −1, 0 −2, 0
Player 1 Middle −1, 0 15, 15 −4, 0
Down −1, −1 11, −2 −2, 0

And now Middle/Centre in round 1 is indeed part of a Nash equilibrium of the overall
two-round game (although this equilibrium is not unique).