Chapter 3

Part 2:
-Bargaining Problems
-Analysis of simple Bargaining Game.
 Introduction
 Contracting is fundamental part in everyday life where parties often negotiate over terms of
engagement

 Example:
 Firm and worker negotiate over whether and how to hire/work together.
 Patient and dentist negotiate over treatment options, cost, and financing.

 Bargaining creates values (dentist earn income, you fix or avoid pain) and divides values (fees,
salary).

 The division depends on social norm, skill, starting positions

 Usually bargaining has a default outcome associated with failing to reach an agreement.
 Representation of Bargaining problems.
 A simple way of mathematically representing a bargaining problem is to describe the
alternatives available to the parties.

 In this topic we will detail all the information, skills, and options available to the two
bargaining parties.

 In other words, the bargaining problems will be focused on interaction of two player.

 The two person bargaining problem consists of a pair (𝑉, 𝑑) where:

 𝑉= sets of 2-vectors payoff reachable by bargaining.
 𝑑= default/disagreement vector.
 Representation of Bargaining problems.
 Example: Two players are trying to decide whether to start a business partnership.
 Suppose that they initiate the partnership, then the business will yield a payoff of 4 for player 1
and payoff of 6 for player 2.
 If they do not form a partnership, then they will each get a payoff of 2.

 Instead of discussing on their negotiation in terms of “form a partnership” and “do not form
a partnership”, we can use the term of end results that is payoff vector (4,6) and (2,2)

 Then 𝑉 = { 4,6 , 2,2 } is called the bargaining set and 𝑑 = (2,2) is the default outcome
that is when they reach disagreement point.
 Representation of Bargaining problems.
 Suppose that 𝑧 represent the non-monetary term. (For example, in the previous example, 𝑧
represent either “forming a partnership” or “do not form partnership”). Then we can say
represent the payoff function as
𝑢𝑖 = 𝑣𝑖 (𝑧)
Where i represent the player and 𝑣𝑖 (𝑧) is a function that maps 𝑧 to some defined value.

 In bargaining, we can say that a monetary transfer is possible to achieve some agreement.
That is over the payoff that player may gain once the outcome reach, they may transfer to
one another some agreeable monetary.
 Representation of Bargaining problems.
 In bargaining problem, we are assuming that monetary transfer are possible that is
 𝑡 = transfer from player 2 to player 1 (if 𝑡 is negative, then it means |𝑡| from 1 to 2)

 Then the payoff function for each player are now can be write as

𝑢1 = 𝑣1 𝑧 + 𝑡, 𝑢2 = 𝑣2 𝑧 − 𝑡

 Such payoff function are called as transferable payoff.

 Representation of Bargaining problems.
 The implication of monetary transfer is that the bargaining set 𝑉 is now can be expanded
that is we can say that if 𝑣 = (𝑣1 , 𝑣2 ) is inside of 𝑉, then 𝑣 = (𝑣1 + 𝑡, 𝑣2 − 𝑡) is also inside
𝑉.

 Example: as the previous example of business partnership

 (2,2) belongs in 𝑉, then (2 + 𝑡, 2 − 𝑡) is inside 𝑉 also.
 (4,6) belongs in 𝑉, then (4 + 𝑡, 6 − 𝑡) is inside 𝑉 also. (ex: let 𝑡 = 2, then (6,4) belongs to 𝑉)

 The sum 𝑢1 + 𝑢2 is called the joint value. It equals to 𝑣1 𝑧 + 𝑣2 𝑧 .

 It is due to transfers affects value division but not the value creation.
 Representation of Bargaining problems.
 Given a cartesian plane where axis is defined by 𝑢1 and 𝑢2 , a fixed joint value is any line
sloped -1

 The graph above shows the bargaining set in the partnership example.
 Representation of Bargaining problems.
 Surplus of an agreement is defined as the difference between the joint value of the contract
and the joint value that would have been obtained had the player not reached an agreement.

 Hence, surplus, 𝑆 is mathematically defined as

𝑆 = 𝑣1 𝑧 + 𝑣2 𝑧 − 𝑑1 − 𝑑2

 A bargaining outcome 𝑣 ∈ 𝑉 is efficient if 𝑢1 + 𝑢2 or joint value is maximized.

 Representation of Bargaining problems.
 The maximized joint value can be intuitively be point out from the illustration graph of
bargaining set of the problem. Example, from the previous problem

 We can see that for the line that passing through the payoff
vector (4,6), the bargaining set contains no point upward and to
the right of the line.

 Then the line indicates the maximized joint value of the

bargaining set.

 Let us say for payoff vector (4,6), Suppose there exist other point in the bargaining set that
is to the right or upward of the point, then the payoff vector is not efficient.
 Rose and Jerry
 Example: in a two-person bargaining problem, let player 1 be Rose that is a principal of
high school and player 2 be Jerry that is a former computer specialist and now a
professional actor.

 Suppose they both agree on Jerry employment into the high school, another aspect need to
be considered by both parties are Jerry responsibilities on the job that is
 Jerry responsible for only drama course
 Jerry responsible for drama course and softball team.

 Given that when Rose and Jerry agree on the employment, Rose get a payoff of 40000$
and Jerry with 10000$.
 Rose and Jerry
 If jerry only assigned to drama course, then he gets the fullness of the 10000$ payoff as he
doesn’t have to put extra effort. Same with Rose.
 If Jerry assigned to drama and softball, his values decreases by 3000$ as the effort need to
expend and Rose gets additional 5000$ if Jerry does.

 In this sense, we can write Rose and Jerry payoff as

𝑣1 𝑥 = 40000 + 5000𝑥, 𝑣2 𝑥 = 10000 − 3000𝑥
Where 𝑥 = 0 indicates responsibilities only on drama
𝑥 = 1 indicates responsibilities on drama and softball.

 If disagreement achieved, the default outcome are (10000,15000) where 𝑑1 = 10000 and
𝑑2 = 15000
 Rose and Jerry

 The joint value are defined by

𝑢1 (𝑥) + 𝑢2 (𝑥) = 𝑣1 𝑥 + 𝑡 + 𝑣2 𝑥 − 𝑡
= 50000 + 2000𝑥
 Hence, the joint value will be maximized when 𝑥 = 1 that is 𝑢1 1 + 𝑢2 1 = 52000
 Rose and Jerry
 Note that as 𝑥 increase from 0 to 1, the payoff for Jerry is decrease. But this can be
compensated by the monetary transfer for example increase in salary.

 The joint value of default outcome is 15000 + 10000 = 25000

 The agreement to set 𝑥 = 1 generates a surplus of

𝑆 = 52000 − 25000 = 27000.
 The standard bargaining solution
 A bargaining solution is a function that assigns to any bargaining problem (𝑉, 𝑑) a feasible
outcome 𝑓 𝑉, 𝑑 ∈ 𝑉.

 Example:
𝑓 = Maximize surplus, give all to player 1
𝑓 = Maximize surplus, give all to player 2
𝑓 = Average surplus, split evenly.
𝑓 = Give each the default payoff
 The standard bargaining solution
 Standard bargaining solution is introduced by John Nash. It is a mathematical
representation of efficiency and proportional division.

 The basic idea is that each player is assumed to obtained his default payoff, plus his share
of payoff.

 Give player 𝑖 a bargaining weight, 𝜋𝑖 ≥ 0 with 𝜋1 + 𝜋2 = 1 that represents powers, skills,

starting position, information,…

 Then we expect the player to reach an agreement in which each player obtains the payoff
𝑢1∗ = 𝑑1 + 𝜋𝑖 𝑆 ∗
 The standard bargaining solution procedure
1. Calculate action 𝑥 ∗ that maximize joint value - call it as 𝑣 ∗

2. Calculate the surplus of the maximized joint value, 𝑆 ∗

3. Given the bargaining weight for each player that is 𝜋1 and 𝜋2 , find each player expected
payoff
𝑢𝑖∗ = 𝑑𝑖 + 𝜋𝑖 𝑆 ∗

4. Calculate 𝑡 such that 𝑢1∗ = 𝑣1 𝑥 ∗ + 𝑡.

5. Hence, for the action 𝑥 ∗ that maximize joint value, monetary transfer 𝑡 ∗ give the standard
bargaining solution of 𝑢1 = 𝑢1∗ , 𝑢2 = 𝑢2∗ with respect the given bargaining weight.
 Rose and Jerry continuation
2 1
 Given the bargaining weight of 𝜋1 = and 𝜋2 = for Rose and Jerry problem. Find the
3 3
standard bargaining solution.
Solution

For the given problem, the action that maximized the joint probability is 𝑥 = 0. Based on the
action, the surplus of the joint value are given by
𝑆 = 27000
Hence by standard bargaining solution, each player payoff is
2
𝑢1 = 𝑑1 + 𝜋1 𝑆 = 10000 + 27000 = 28000
3
1
𝑢2 = 𝑑2 + 𝜋2 𝑆 = 15000 + 27000 = 24000
3
 Rose and Jerry continuation
Conti.

We know that
𝑢1 = 𝑣 1 + 𝑡
Since 𝑢1 = 28000 is the payoff for Rose, then the monetary transfer, 𝑡 needed to achieve is
28000 = 40000 + 5000 1 + 𝑡
28000 = 45000 + 𝑡
𝑡 = −17000

Hence, the best action that maximize the joint value is 𝑥 = 1 and monetary transfer 𝑡 =
− 17000 to obtained the standard bargaining solution 𝑢1 = 28000, 𝑢2 = 24000 with the
given bargaining weight 𝜋1 and 𝜋2 .
 In class Exercise:
Use the standard bargaining solution to find the outcomes of the following bargaining
problems. Player 1’s payoff is 𝑢1 = 𝑣1 (𝑥) + 𝑡 and player 2’s payoff is 𝑢2 = 𝑣2 (𝑥) − 𝑡.
Graph the maximized joint value and the default outcome; report the chosen 𝑥, 𝑡 as well as the
players’ individual payoffs.
 Exercise:
Use the standard bargaining solution to find the outcomes of the following bargaining
problems. Player 1’s payoff is 𝑢1 = 𝑣1 (𝑥) + 𝑡 and player 2’s payoff is 𝑢2 = 𝑣2 (𝑥) − 𝑡. In
each case, Graph the maximized joint value and the default outcome; report the chosen 𝑥, 𝑡 as
well as the players’ individual payoffs.

a.

b.
Analysis of simple Bargaining
Problem
 Introduction.
 In contracting, a negotiation can happen multiple times until an agreement can be achieved.

 In the previous subtopic, we have developed some of the basic concepts and language for
study of bargaining problems.

 The exercise allows us to generate intuition about the attainment of efficient outcomes and
about the determinants of bargaining power (bargaining weight).

 In this subtopic we will discuss on how player’s patience and his ability to make offers at
strategically important points in time impact on the bargaining strength.
 Insight
 To get a picture of bargaining strength, we continue discussing the problem on Rose and
Jerry.

 From the previous problem, both achieve a standard bargaining solution

𝑢 = (28000,2400)
2 1
with bargaining weight 𝜋1 = and 𝜋2 = .
3 3

 Note that, suppose that 1 of them has a bargaining weight of 0 for example 𝜋1 = 0. The
bargaining problem still can achieve optimal solution that is standard bargaining solution
𝑢1 = 10000 + 0 27000 = 10000
𝑢2 = 15000 + 1 27000 = 42000
 Insight.
 The outcome is still more efficient compared to the default outcome that is
𝑢 = (10000,15000)

 Now, if a player can decide on determination such bargaining weight, then it only rational
that he decide the bargaining weight to all fall in his favor.

 So, suppose that Rose making an offer of this employment contract, she should set her
bargaining weight equal to one so that her payoff will be optimal.

 For Jerry, any positive value of bargaining weight should be accepted since the bargaining
weight defining only the portion of surplus he may get, and he already has a minimum of
default payoff as his outcome.
 Ultimatum Game – One Round
 In deciding the bargaining weight 0 < 𝜋𝑖 < 1, the bargaining process can be illustrated by
1 simple infinite strategy game that is Ultimatum Game.
 Players: Proposer 𝑖, and Recipient 𝑗.
 Strategy: Given $1 to be split, 𝑖 proposes 𝑚 ∈ 0,1 = 𝑠𝑖 to be given to 𝑗; 𝑗 decides to either
𝑑 𝑚 = 𝐴, 𝑅 that is accept or reject 𝑚
 Payoffs: If proposal rejected, both get (0,0); if accepted, that is 𝑑 𝑚 = 𝐴, then j gets 𝑚, and 𝑖
gets 1 − 𝑚
 Recipient’s two optimal strategies
 How does recipient optimally respond to proposal to get 𝑚?
 Suppose 𝑚 > 0, Recipient gets 𝑚 by accepting, 0 by rejecting → Optimal to accept
 Suppose m=0, Then recipient’s payoff of accepting and rejecting are the same
→ Optimal to either accept or reject.

 Strategy for recipient is a decision function 𝑑 = 𝑑(𝑚) that is 0,1 → 𝐴, 𝑅

 Conclusion: Recipient has two optimal strategies:

 𝑑 𝑚 = 𝐴 always (i.e. accept any proposal)
 𝑑 𝑚 = 𝐴 if 𝑚 > 0, 𝑅 if 𝑚 = 0 (rejects only 𝑚 = 0)

 Later is vindictive at 𝑚 = 0: it has proposer gets $0, instead of $1 even though recipient
gain nothing
 Backward induction-SPE equilibria
 How does a proposer optimally propose, while anticipating recipient’s strategy/overall
subgame?

 Suppose recipients accepts any proposal

 Then proposer get 1-m & optimize by proposing m=0. So, one BI/SPE is profile
𝑠𝑖 , 𝑠𝑗 = 𝑚 = 0, 𝑎𝑐𝑐𝑒𝑝𝑡 𝑎𝑛𝑦 𝑝𝑟𝑜𝑝𝑜𝑠𝑎𝑙
 Suppose recipients is vindictive
 Then proposer gets 0 by proposing m=0, but 1-m>0 by proposing m>0. but the optimal
needs m to be the lowest positive number which does not exist! So, there is no BI/SPE
with vindictive recipient.

 Conclusion: The ultimatum game has a unique BI/SPE, namely

𝑚, 𝑑 = 0, 𝑎𝑐𝑐𝑒𝑝𝑡 𝑎𝑛𝑦 𝑝𝑟𝑜𝑝𝑜𝑠𝑎𝑙 with payoff ($1,0)
 Alternating-offer game.
 As you can see, the one who make offer has the bargaining strength that favoring his
payoff or one with power of ultimatum wins all surplus (𝜋𝑖 = 1)

 In Rose and Jerry example, the offer are non-negotiable. It means the offer will stop on
Jerry decision on accepting the offer or not. This give Rose the power of ultimatum.

 In many real settings, each sides were alternately making offer until one is accepted.

 For example, in the previous problem, Jerry could make a counteroffer which now gives
him the bargaining strength to favoring his payoff.
 Alternating-offer game.
 We can see that whenever a player gets to make offer, he holds the bargaining strength
favoring his own payoff.

 Although alternating offer can be infinite, we need to know that offers and counteroffer
sometimes takes time.

 For example, in home sales, one party may wait a week or more for an offer to be
considered and a counteroffer returned.

 As a saying goes, “Time is Money” which means that most people are impatient to some
degree that is they prefer not to endure protracted negotiation procedure.
 Alternating-offer game: Discount factor
 Most people discount the future relative to the present.

 Example: Suppose there is a seller that sell clothes at a price of RM100. A customer came
and make offer to buy it for RM80. Instead of fixing the payoff(outcome) to RM100 that
leads multiple alternate-offer afterwards, the seller may just discount it and take the offer.

 How a person discounts the future may affect his or her bargaining

 To incorporate the discounting into game-theoretic models, we use the notion of a discount
factor.
 Alternating-offer game: Discount factor
 A discount factor 𝛿𝑖 for player 𝑖 is a number used to deflate the payoff received tomorrow
so that it can be compared with a payoff received today.

 To capture that player’s impatient toward long negotiation (each alternation being one
waiting period) let them discounts future payoffs at the rates 0 < 𝛿1 , 𝛿2 < 1.

 The discount factor 𝛿 is a measure of impatience.

 If the player is patient, then 𝛿𝑖 ~1.

 Alternating-offer: two rounds
 Given $1 to be split, player 1 proposes that 2 keep 𝑚1 ∈ 0,1 ; player 2 can accept for
payoffs (1 − 𝑚1 , 𝑚1 ), or reject and then make a counteroffer. Player 2 proposes that player
1 keeps 𝑚2 ∈ 0,1 ; player 1 can accepts for payoffs (𝛿1 𝑚2 , 𝛿2 1 − 𝑚2 , or reject for
payoffs (0,0)

 Note
 𝑚 is a proposal of what other keep
 Future cash flow is discounted by impatience
 The last round is a subgame identical to ultimatum game
 SPE of 2-round version
 SPE: A profile of strategies that defines a NE is every subgame.

 Focus on the last subgame, where 2 is an “ultimatum proposer”. Its continuation value is
(0,1), which is discounted to (0, 𝛿2 ) in terms of first period payoffs.

 How should 2 optimally decide to 1’s proposal 𝑚1 ?

 By rejecting, her payoff is 𝛿2 ; by accepting, it is 𝑚1 ? So:

 It is optimal to accept if 𝑚1 > 𝛿2 ,
 It is optimal to reject if 𝑚1 < 𝛿2 ,
 It is optimal to either if 𝑚1 = 𝛿2
 SPE of 2-round version
 Again, this indifference defines two strategy for player 2, but “vindictive one” cannot be
optimized against by player 1, hence cannot be part of a SPE

 So, player 1 optimizes by proposing m1 = 𝛿2 , to be accepted.

 Hence, the SPE: (𝑚1 = 𝛿2 & 𝑎𝑐𝑐𝑒𝑝𝑡 𝑎𝑙𝑙, 𝑎𝑐𝑐𝑒𝑝𝑡 𝑖𝑓𝑓 𝑚1 ≥ 𝛿2 & 𝑝𝑟𝑜𝑝𝑜𝑠𝑒 𝑚2 = 0)

 The equilibrium payoffs turn out to be (1 − 𝛿2 , 𝛿2 ),

 Equilibrium involves immediate acceptance in round 1.

 SPE of 2-round version
 Example, suppose that in the Rose and Jerry problem, Jerry get to make a counteroffer.
Which is to reject proposal by Rose and then make an offer. Given that the school needs
staff as soon as possible and Jerry need the work as soon as possible too, both discounts the
future by 𝛿1 = 0.8, 𝛿2 = 0.7.

 Solution:
Since this negotiation involves with 2 round offer, the ultimatum will go to player 2 that is
jerry. The equilibrium payoff for of this 2 round offer game is
1 − 𝛿2 , 𝛿2 = (0.3,0.7)

Hence, for Rose to make strategy to optimal the payoff, she should give a bargaining weight
of 𝜋1 = 0.3 and 𝜋2 = 0.7 so that the game may ends at round 1 only.
 Alternating offer 3-round version
 Given $1 to be split, player 1 proposes that 2 keep 𝑚1 ∈ 0,1 ; player 2 can accept for
payoffs (1 − 𝑚1 , 𝑚1 ), or reject and then make a counteroffer.
Player 2 proposes that player 1 keeps 𝑚2 ∈ 0,1 ; player 1 can accepts for payoffs
(𝛿1 𝑚2 , 𝛿2 1 − 𝑚2 , or reject and then continue making a counteroffer.
Player 1 then proposes that player 1 keeps 𝑛1 ∈ 0,1 ; player 1 can accepts for payoffs
(𝛿12 (1 − 𝑛1 ), 𝛿22 𝑛1 , or reject for payoff.

 After round 1, subgame = 2-round game with role reversed continuation value is
𝛿1 , 1 − 𝛿1 . (previous slide, roles reversed)

 In round 1, 2’s optimal strategy that is part of SPE is accept iff 𝑚1 ≥ 𝛿2 1 − 𝛿1 . Player 1
optimizes with 𝑚1 = 𝛿2 (1 − 𝛿1 )

 Equilibrium payoff turn out to be (1 − 𝛿2 1 − 𝛿1 , 𝛿2 1 − 𝛿1 )

 Alternating offer 3-round version
 Continue with Rose and Jerry, Rose get to reject the counteroffer and continue to make
offer.

 The bargaining problem are now became a 3 round offer game. Hence the equilibrium
payoff are now
1 − 𝛿2 1 − 𝛿1 , 𝛿2 1 − 𝛿1 = 0.86,0.14

 Means that the bargaining weight are now 𝜋1 = 0.86 and 𝜋2 = 0.14.
 Whole Game SPE
 In determining the whole game SPE, we need to take a look at each stage of bargaining by
working backwards.

 Example: Consider the following bargaining problem: $20 dollars needs to be split between
Jack and Jill. Jill gets to make an initial offer. Jack then gets to respond by either accepting
Jill’s initial offer or offering a counteroffer. Finally, Jill can respond by either accepting
Jakes offer or making a final offer. If Jack does not accept Jill’s final offer both Jack and
Jill get nothing. Jack discounts the future at 10% (i.e. future earnings are with 10% less
than current earnings while Jill discounts the future at 20%. Calculate the Nash equilibrium
of this bargaining problem.
 Whole Game SPE
 Solution:
At any stage, the offer made needs to be acceptable to both parties. We need to work
backwards:
Stage 3: Note that if Jack rejects Jill’s offer at this stage, the money disappears. Therefore,
Jack will accept anything positive.

Jill offers: $20 to herself, $0 to Jack

Stage 2: Now, Jack must make an offer that Jill will accept (if the game gets to stage three,
Jack gets nothing). Jill is indifferent between $20 in one year and $16 today (she discounts the
future at 20%).

Jack offers: $16 to Jill, $4 to Himself

 Whole Game SPE
Stage 1: Now, Jill must make an offer that Jack will accept (and is preferable to her – if this is
not possible, then she will make an offer jack rejects and the game goes to stage 2). Jack is
indifferent between $4 in one year and $3.60 today (he discounts the future at 10%). Note that
$16.40 is preferred by Jill to $16 in one year.

Jill offers: $16.40 to herself, $3.60 to Jack

 Whole Game SPE
Exercise: $40 dollars needs to be split between Jack and Jill. Jill gets to make an initial offer.
Jack then gets to respond by either accepting Jill’s initial offer or offering a counter offer.
Finally, Jill can respond by either accepting Jack offer or making a final offer. If Jake does not
accept Jill’s final offer both Jack and Jill get nothing. Jack discounts the future at 15% (i.e.
future earnings are with 15% less than current earnings while Jill discounts the future at 10%.
Calculate the Nash equilibrium of this bargaining problem.