Page |1

Game Theory
Game theory is a branch of economics that studies strategic interactions among rational decision-
makers. It provides a framework for analyzing how individuals, firms, or other economic agents make
decisions in situations where their outcomes depend not only on their own choices but also on the
choices of others.

Elements / Ingredients of a Game: Generally, a Game has four elements.

1. Players: A game must have two or more players.

2. Strategies / Interactions: Each player can have multiple choices; players can choose any one of
them or can impose probabilities to each choice. These are called strategies.
There have different types of strategies. Some of them are,
i) Dominant Strategy
ii) Pure Strategy
iii) Mixed Strategy

Dominant Strategy: A dominant strategy is a strategy that provides the highest payoff for a
player, regardless of the strategies chosen by other players. In other words, a dominant
strategy is the best course of action for a player, regardless of what their opponents do.

Pure Strategy: When a player chooses any one of his options without assigning any probabilities
and sticking to it, then it is called pure strategy.

Example: In a simple example, consider a game where two players can choose between "confess"
or "Deny." If Player A decides to always confess, regardless of what Player B does, A is using a
pure strategy.

Mixed Strategy: When a player plays every choice a portion of time or assign probabilities to
his choices then it is called mixed strategy. A mixed strategy is a strategy that involves a player
randomizing between two or more pure strategies with a certain probability distribution.

3. Payoff: In a game every choice must have payoffs for players. The payoff of the players is
represented by payoff matrix. In every cell of a payoff matrix the first outcome represents
the payoff of row and the second outcome represents the payoff of column.
4. Rules and Regulations: Every player have to follow some criteria.

The Payoff Matrix of a Game

© Md Khaledur Rahman
 Page |2

Payoff matrix is a table that shows the outcomes of different choices of players in a game. In a payoff
matrix the first outcomes in every cell represents the outcome of rows and the second outcome
represents the outcome of column.

A payoff matrix of a game

Player B
Left Right
Top 1, 2 0, 1
Player A
Bottom 2, 1 1, 0
In the payoff matrix the black outcomes are the payoff of rows and the blue ones are the payoff of
column.

Nash Equilibrium
A pair of strategy is Nash equilibrium if A’s choice is optimal given B’s choice and B’s choice is optimal
given A’s choice.

A payoff matrix of a game

Player B
Left Right
Top 1, 2 0, 1
Player A
Bottom 2, 1 1, 0
In the above payoff matrix suppose player A choose top, then Player B will choose left as it provide
highest payoff. Again, the strategy will be Nash equilibrium if at the same time Player-A choose top
given Player-B choose left. But in the above payoff matrix given Player B choose left, Player-A will
choose bottom as it pays highest payoff. So, it is not a Nash equilibrium.

Similarly, if player A choose bottom, player B will choose left. At the given player B choose left, player
A will choose bottom. So, it is a Nash equilibrium.

A payoff matrix of a game

Player B
Left Right
Top 1, 2 0, 1
Player A
Bottom 2, 1 1, 0

The circle pointed strategy is Nash equilibrium.

A payoff matrix can have multiple Nash equilibrium or no Nash equilibrium.

© Md Khaledur Rahman
 Page |3

Player B
Left Right
Top 2, 1 0, 0
Player A
Bottom 0, 0 1, 2
In the above payoff matrix, there have two Nash equilibrium.

A game with no Nash equilibrium (in pure strategies)

Player B
Left Right
Top 0,0 0, -1
Player A
Bottom 1,0 -1, 3
And in this payoff matrix there have no Nash equilibrium.

Mixed Strategy

In pure strategy player choose any one strategy from different choices. But in mixed strategy each
player assigns probabilities to each choice and play the game according to the probability.

The Prisoner’s Dilemma

Nash equilibrium does not always provide Pareto efficient outcomes.

For example, we consider two prisoners who had committed a crime together. Both of them are being
questioned in two separate room about the crime. If both prisoners confess, they both will be punished
3-month jail. If prisoner A confess and prisoner B deny, prisoner A will be free and prisoner B will be
punished 6 months jail. Reverse case will happen if prisoner A deny and prisoner B confess. But, if both
of them deny, they will be punished only 1 month jail.

Prisoner B
Confess Deny
Confess -3, -3 0, -6
Prisoner A
Deny -6, 0 -1, -1

Given prisoner A confess, prisoner B will confess too. Because it provides him maximum payoff.
Similarly, if prisoner B confess, prisoner A confess too. So, it is a Nash Equilibrium. Again, if prisoner
A deny, prisoner B will confess because he has not to be punished in this strategy rather if he denies
too, he has to be punished 1 month jail. At the same time, if prisoner B confess, the prisoner A will
confess otherwise he have to be punished 6 months jail. So, it is not Nash Equilibrium. In the payoff
matrix there have only one Nash equilibrium (-3, -3).

But it is not Pareto efficient outcome. Because, in the equilibrium both of them will be punished 3
months jail. But they could be punished only 1 month without hurting each other if they both deny. So,

© Md Khaledur Rahman
 Page |4

(deny, deny) would be Pareto efficient outcome. But it is only possible when cooperation is available
between them.

In the prisoner’s dilemma if the game occurs for definite number of times, then the prisoners never
go to cooperation. But, if it occurs for indefinite number of times, they will cooperate.

Sequential Game
In a sequential game one of the players chooses first and after seeing his choice the other player
chooses. The first person who chooses first is call “leader” and the second player is known as “follower”.

Incumbent
Fight Don’t Fight
Stay Out 1, 9 1, 9
Entrant
Stay In 0, 0 2, 1

1, 9

1, 9

0, 0

2, 1

In the first row two outcomes are same because if entrant stay out there is no need to fight and the
outcomes remains same.

In a sequential game incumbent can threat the entrant not enter into market. The incumbent can threat
the entrant that if he enters the market incumbent will fight and the outcome will become zero for
entrant though he getting outcome 1 if he doesn’t enter the market.

There are two types of threat.

 Credible (িব াসেযাগ ) threat

 Incredible (অিব াসেযাগ ) threat.

In the given payoff matrix, the threat given by incumbent is incredible. Because, if he fights, his
outcome decreases too [if he doesn’t fight, he will get 1. But if he fights, he will get 0].

© Md Khaledur Rahman
 Page |5

But it can be credible if the outcome for incumbent can be increased if he fights. In this case-

Incumbent
Fight Don’t Fight
Stay Out 1, 9 1, 9
Entrant
Stay In 0, 2 2, 1

In this case the threat would be credible.

Game Applications
The Best Response Curves
Best response is the response, given other players choice, that will maximize my payoff.

Column
Left Right
Top 2, 1 0, 0
Row
Bottom 0, 0 1, 2

In the payoff matrix best responses are:

Row chooses: Top Bottom

Column’s best response: Left Right

Column chooses: Left Right

Row’s best response: Top Bottom

Because, these pay highest payoff to any player given other players choice.

In a pure strategy all Nash equilibrium is best response for both players. But, all best response for a
player is not Nash equilibrium.

Mixed Strategies

© Md Khaledur Rahman
 Page |6

Mixed strategy is the strategy where each players assign probabilities to each choice and plays the
game according to that probability.

Ms. Column
Left (c) Right (1-c)
Top (r) 2, 1 0, 0
Mr. Row
Bottom (1-r) 0, 0 1, 2
r is the probability that row plays top, and (1 − r) is the probability that he plays bottom. Similarly, c
is the probability that column plays left, and (1−c) is the probability that she plays right.

So, payoff of Mr. Row for different combinations:

Combination Probability Payoff

Top, Left rc 2
Top, Right (1-r)c 0
Bottom, Left r(1-c) 0
Bottom, Right (1-r)(1-c) 1

Expected Payoff of Mr. Row = 2𝑟𝑐 + 1 − 𝑟 − 𝑐 + 𝑟𝑐 = 3𝑟𝑐 − 𝑟 − 𝑐 + 1

If Mr. Row want to change his probability of choices, then the change in payoff is expressed as—

Δ𝑝𝑎𝑦𝑜𝑓𝑓 𝑜𝑓 𝑟𝑜𝑤 = 3Δ𝑟𝑐 − Δr

= (3𝑐 − 1)Δ𝑟

Mr. Row will increase r when, Mr. Row will decrease r when, Mr. Row will remain unchanged,
3𝑐 − 1 > 0 3𝑐 − 1 < 0 3𝑐 − 1 = 0
3𝑐 > 1 3𝑑 < 1 3𝑑 = 1
1 1 1
𝑐> 𝑐< 𝑐=
3 3 3

Now, payoff of Ms. Column for different combinations:

Combination Probability Payoff

Left, Top cr 1
Right, Top c(1-r) 0
Left, Bottom (1-c) r 0
Right, Bottom (1-c)(1-r) 2

Expected Payoff of Mr. Row = 𝑐𝑟 + 2 − 2𝑐 − 2𝑟 + 2𝑐𝑟 = 3𝑐𝑟 − 2𝑐 − 2𝑟 + 2

If Ms. Column want to change her probability of choices, then the change in payoff is expressed as—

Δ𝑝𝑎𝑦𝑜𝑓𝑓 𝑜𝑓 𝑐𝑜𝑙𝑢𝑚𝑛 = 3Δcr − 2Δc

= (3𝑟 − 2)Δ𝑐

© Md Khaledur Rahman
 Page |7

Ms. Column will increase r Ms. Column will decrease r Ms. Column will remain unchanged
when, when, when,
3𝑟 − 2 > 0 3𝑟 − 2 < 0 3𝑟 − 2 = 0
3𝑟 > 2 3𝑟 < 2 3𝑟 = 2
2 2 2
𝑟> 𝑟< 𝑟=
3 3 3

1 C 1/3 0

C r

1/3 2/3

0 1
r 2/3

Figure: Best Response Curve

In this game we have three Nash equilibrium. These are , , (0, 0) 𝑎𝑛𝑑 (1, 1)

Game of Coordination
Battle of Sexes
In this game a boy and a girl want to meet at a movie but they do not have phones to coordinate with
each other. The boy likes action film and the girl likes art film. So, their payoff matrix is –

The battle of the sexes

Girl
Action Art
Action 2, 1 0, 0
Boy
Art 0, 0 1, 2

© Md Khaledur Rahman
 Page |8

In this game they can meet of both of them goes to same movie. So, the Nash equilibrium is (Action,
Action) and (Art, Art).

But, as they like different movies, how can they meet? How will one know what the other is choosing?

If they have phones or they can coordinate by any way, they can meet at same movie. Or, if any one
movie is near of their home, then both of them can think to go to the neared movie. Without these
reasons they cannot reach at equilibrium as they like different movies.

Prisoner’s Dilemma
In a prisoner’s dilemma “confess” is the dominant strategy though coordination is possible or not
possible. But it is not Pareto efficient.

But they can reach at Pareto efficient outcome if coordination between them is available and the game
repeats indefinitely.

Assurance Game
Arms Race
In this game both U.S. and U.S.S.R. have two strategies, build nuclear weapon or refrain from building
nuclear weapon. So, in this game their payoffs are –

An arms race

U.S.S.R.
Refrain Build
Refrain 4, 4 1, 3
U.S.
Build 3, 1 2, 2
In this game there is two pure strategy Nash equilibrium (refrain, refrain) and (build, build). If both
of them refrain, they will get the highest payoff. But the problem is that neither party knows which
choice the other will make. So, before committing refrain, each party wants some assurance that the
other will refrain. In this game if one player moves first to refrain then they can reach the equilibrium.

© Md Khaledur Rahman
 Page |9

Game of Competition
Zero-Sum Game
In a zero-sum game the payoff of one player is equal to the losses of the other player. For example,
we consider a Penalty in Soccer. For this game the payoff matrix is ---

Penalty points in soccer

Column
Defend left Defend right
Kick left 50, -50 80, -80
Row
Kick right 90, -90 20, -20
The game is considered in mixed strategy because the players divide their kicks and defends in both
side in spite of kicking or defending on one side.

Monopolistic Competition
There are two types of games:

 Simultaneous Game
 Sequential Game

Simultaneous Game:
(Cournot Model) There are two firms in a market, firm 1 and 2. Firm 𝑖 = 1 𝑜𝑟 2 has a linear cost
function 𝑐(𝑞 ) = 2𝑞 . The market price is determined by 𝑝 = 20 − 2𝑞 − 3𝑞 . The two firms
simultaneously choose their quantities. Solve for the Nash Equilibrium of (𝑞 , 𝑞 ).

Answer: The two firms’ profit functions are:

𝜋 = 𝑝𝑞 − 𝑐𝑞 = (𝑝 − 𝑐)𝑞 = (20 − 2𝑞 − 3𝑞 − 2)𝑞

𝜋 = 𝑝𝑞 − 𝑐𝑞 = (𝑝 − 𝑐)𝑞 = (20 − 2𝑞 − 3𝑞 − 2)𝑞

Let’s derive each firm’s best response function.

Firm 1:

𝜋 = 𝑝𝑞 − 𝑐𝑞
= (𝑝 − 𝑐)𝑞
= (20 − 2𝑞 − 3𝑞 − 2)𝑞
= (18 − 2𝑞 − 3𝑞 )𝑞
= 18𝑞 − 2𝑞 − 3𝑞 𝑞

© Md Khaledur Rahman
 P a g e | 10

To find the best response we differentiate them,

Thus, Firm1’s best response function is

𝜕𝜋
⇒ 18 − 4𝑞 − 3𝑞 = 0
𝜕𝑞
⇒ 18 − 3𝑞 = 4𝑞
18 − 3𝑞
𝑞 =
4
That means at that quantity (𝑞 ), firm 1 will maximize profit.

Firm 2:

𝜋 = 𝑝𝑞 − 𝑐𝑞
= (𝑝 − 𝑐)𝑞
= (20 − 2𝑞 − 3𝑞 − 2)𝑞
= (18 − 2𝑞 − 3𝑞 )𝑞
= 18𝑞 − 2𝑞 𝑞 − 3𝑞

To find the best response we differentiate them,

Firm 2’s best response function is

𝜕𝜋
= 18 − 2𝑞 − 6𝑞 = 0
𝜕𝑞
18 − 2𝑞
𝑞 =
6

The N.E of (𝑞 , 𝑞 ) should satisfy the best response function of both firms.

18 − 2𝑞1
𝑞 =
6
18
4
N.E
3 18 − 3𝑞2
𝑞 =
4

2 3 6

We can solve the best response functions

© Md Khaledur Rahman
 P a g e | 11

18 − 4𝑞 − 3𝑞 = 0
18 − 2𝑞 − 6𝑞 = 0

and get the N.E of (𝑞 , 𝑞 ) is (𝑞 = 3, 𝑞 = 2) .

Sequential Game:
(Stackelberg Game) There are two firms in a market, firm 1 and 2. Firm 𝑖 = 1 𝑜𝑟 2 has a linear cost
function 𝑐(𝑞 ) = 20𝑞 . The market price is determined by 𝑝 = 100 − 4(𝑞 + 𝑞 ).

Suppose the two firms move sequentially to choose their quantities: firm 1 chooses 𝑞 at period 1, and
then firm 2 chooses 𝑞 at period 2. Solve for the subgame perfect Nash Equilibrium of (𝑞 , 𝑞 ) and the
profit of the two firms.

Backward Induction:

Solution: we solve the sequential game using backward induction.

Period 2:

𝜋 = 𝑝𝑞 − 𝑐𝑞
= (𝑝 − 𝑐)𝑞
= (100 − 4𝑞 − 4𝑞 − 20)𝑞
= 80𝑞 − 4𝑞 𝑞 − 4𝑞

Firm 2’s best response function is

𝜕𝜋
⇒ 80 − 4𝑞 − 8𝑞 = 0
𝜕𝑞
⇒ 8𝑞 = 80 − 4𝑞
80 − 4𝑞
⇒𝑞 =
8
20 − 𝑞
⇒𝑞 =
2

Period 1:

𝜋 = 𝑝𝑞 − 𝑐𝑞
= (𝑝 − 𝑐)𝑞
= (100 − 4𝑞 − 4𝑞 − 20)𝑞
= 80𝑞 − 4𝑞 − 4𝑞 𝑞
20 − 𝑞
= 80𝑞 − 4𝑞 − 4𝑞
2
160𝑞 − 8𝑞 − 80𝑞 + 4𝑞
=
2
80𝑞 − 4𝑞
=
2

© Md Khaledur Rahman
 P a g e | 12

= 40𝑞 − 2𝑞

Firm 1’s best response function is

∂π
⇒ 40 − 4𝑞 = 0
𝜕𝑞
⇒ 4𝑞 = 40
⇒ 𝑞 = 10
So,

20 − 𝑞
𝑞 =
2
20 − 10
=
2
=5

Thus, the subgame perfect Nash equilibrium is (𝑞 , 𝑞 ) = (10, 5)

Note: For simultaneous game we find Nash equilibrium and for sequential game we find Subgame
perfect Nash equilibrium.

Oligopoly
If there is only one seller in a market, it is called a monopoly market. The number of sellers is two in a
duopoly market. But, in an oligopoly market there is few sellers.

To see the interaction of firms in an oligopolistic market we obtain, for simplicity, there are two firms
and both of them produces identical product and each firm decides to set the price of his goods and
set the profit maximizing quantity. When one firm decides about its choices for price and quantities it
may already know the choices made by the other firm. The firm who set his price or quantity first is
called price or quantity leader and the other firms are called the price or quantity follower. This is a
sequential game because one chooses first and the other choose after the leader chooses.

In the sequential game the equilibrium for quantity leadership is call Stackelberg Equilibrium.

Here, total output is the sum of the output of the two firms. Thus,

𝑌 = 𝑦 +𝑦

And price,

𝑝(𝑦 + 𝑦 ) = 𝑎 − 𝑏(𝑦 + 𝑦 )

For convenience we will take costs to be zero. So, the MC is zero too.

© Md Khaledur Rahman
 P a g e | 13

The Follower’s Problem:

The follower takes the quantity produced by leader as constant. So, the follower will choose the
quantity to produce where his marginal cost and marginal revenue is equal.

𝑀𝑅 = 𝑀𝐶

Thus, followers profit maximization—

max 𝜋 = 𝑝(𝑦 + 𝑦 )𝑦 − 𝑐 (𝑦 )

Here,

𝑇𝑅 = 𝑝(𝑦 + 𝑦 )𝑦
= [𝑎 − 𝑏(𝑦 + 𝑦 )]𝑦
= 𝑎𝑦 − 𝑏𝑦 𝑦 − 𝑏𝑦

𝜕𝑇𝑅
⇒ 𝑎 − 𝑏𝑦 − 2𝑏𝑦 = 0
𝜕𝑦
⇒ 2𝑏𝑦 = 𝑎 − 𝑏𝑦
𝑎 − 𝑏𝑦
⇒𝑦 =
2𝑏

The Leader’s Problem:

Leader’s profit maximization—

max 𝜋 = 𝑝(𝑦 + 𝑦 )𝑦 − 𝑐 (𝑦 )

Here,

𝑇𝑅 = 𝑝(𝑦 + 𝑦 )𝑦
= [𝑎 − 𝑏(𝑦 + 𝑦 )]𝑦
= 𝑎𝑦 − 𝑏𝑦 − 𝑏𝑦 𝑦
𝑎 − 𝑏𝑦
= 𝑎𝑦 − 𝑏𝑦 − 𝑏𝑦
2𝑏
𝑎 − 𝑏𝑦
= 𝑎𝑦 − 𝑏𝑦 − 𝑦
2
2𝑎𝑦 − 2𝑏𝑦 − 𝑎𝑦 + 𝑏𝑦
=
2
𝑎𝑦 − 𝑏𝑦 1
= = (𝑎𝑦 − 𝑏𝑦 )
2 2

𝜕𝑇𝑅 1
⇒ (𝑎 − 2𝑏𝑦 ) = 0
𝜕𝑦 2
⇒ 𝑎 − 2𝑏𝑦 = 0 [ 𝑚𝑢𝑙𝑡𝑖𝑝𝑙𝑖𝑛𝑔 𝑏𝑜𝑡ℎ 𝑏𝑦 2]

© Md Khaledur Rahman
 P a g e | 14

⇒ 2𝑏𝑦 = 𝑎
𝑎
⇒𝑦 =
2𝑏

So,
𝑎 − 𝑏𝑦
𝑦 =
2𝑏
𝑎 𝑎
𝑎−𝑏 𝑎−2
= 2𝑏 =
2𝑏 2𝑏
2𝑎 − 𝑎
2 𝑎
= =
2𝑏 2 × 2𝑏
𝑎
=
4𝑏

Externalities
When anyone’s consumption or production affects others positively or negatively and for which no
payment or compensation is made then the situation is called externalities.

There can be two types of externalities:

1. Positive Externalities
2. Negative Externalities

Positive Externalities: When anyone’s activity affect others positively and for which no payment is
made then the situation is called positive externalities.

For example, if my neighbor makes a flower garden in front of her house, it intensifies the beauty of
my house and also gives me good smells. But I do not pay her anything for this.

Negative Externalities: When anyone’s activity affect others negatively and for which no
compensation is made then the situation is called negative externalities.

For example, if my roommate smokes inside of room, though I don’t like smoking, and for that he does
not pay me any compensation. This is an example of negative externalities.

There is another type of externalities and that is known as Mutually Positive Externalities.

Mutually Positive Externalities: When two person becomes affected positively by the activity of each
other and for which no payment is made, then the situation is called mutually positive externalities.

© Md Khaledur Rahman
 P a g e | 15

For example, if I cultivate fruit trees and my neighbor do beekeeping then his beekeeping helps in the
pollination of my trees and the bee collect honey from my garden that his honey production and both
becomes positively affected.

Note: there is no market for externalities to trade.

Public Good
Public good is defined by two characteristics. These are---

1. Non-excludable
2. Non-rival

Non-excludable: No one can be excluded from the consumption of the public goods.

Non-rival: Consumption of one will not reduce the consumption of others.

Tragedy of Commons: When property right is not defined then there is a possibility that the property
will be over used, then tragedy of commons happens.

To solve the tragedy of commons some steps should be taken.

 Property Right of public goods should be declared.

 Public goods should be provided under the privet observation.

Exchange
Partial Equilibrium: If the equilibrium for only one good is considered then it is called partial
equilibrium.

For example: In a market there are several goods, chicken is one of them. If we consider the supply
and demand of chicken in an equilibrium model then it is called the partial equilibrium.

General equilibrium: Considering all goods in a market system the equilibrium we get is called general
equilibrium.

For example: Mutton is a substitute of chicken. If the price of mutton decreases much the demand of
chicken will be affected. Like this there have much substitute and complements of a good in the market.
Considering the price of all, we find general equilibrium.

© Md Khaledur Rahman
 P a g e | 16

To analyze a general equilibrium problem, we simplify the model as follows,

1. We take the market as perfect competitive.

2. We consider only two goods and two consumers.
3. We fix the endowment (supply), no production will be involved, and examine how they trade these
goods among themselves.

To show general equilibrium we use Edgeworth box.

Edgeworth Box: Edgeworth box is a graphical tool which is used to analyst the exchange of two goods
(Good 1 and Good 2) between two consumers (Consumer A and Consumer B).

Feasible Allocation: An allocation is feasible allocation when the total amount of goods consumed is
equal to the total amount of goods available (endowment). Thus,

𝑋 +𝑋 =𝜔 +𝜔

𝑋 +𝑋 =𝜔 +𝜔

In an Edgeworth box, the width of the box measures the total amount of good 1 (sum of the consumption
of good-1 for consumer 1 & 2) in the economy and the height measures the total amount of good 2.

The points in the box give us both the bundles that A can hold and the bundles that B can hold and
these points represent the feasible allocations in the economy. Person A’s consumption choices are
measured from the lower left-hand corner while person B’s choices are measured from the upper right.
In an Edgeworth box we can see the preferred consumption bundles of both consumers.

[Note: Blue one is the IC of Person-A and the black one is the IC of Person-B]

Trade:

We start at an initial endowment point denoted by ‘W’ where both consumers choose their consumption
bundles. This is the point where the indifference curves of both consumers intersect. Both of them
will be better off if they can choose any bundle above the current indifference curve and this can be
happened only if both of them choose any bundle (M) in the lens-shaped region where they can reach
higher indifference curve. In this case they will find some mutually advantageous trade that will move

© Md Khaledur Rahman
 P a g e | 17

them to some point inside the lens-shaped area such as the point M. In this trade Consumer-1 will give
up {𝑤 − 𝑥 } units of good-1 and Consumer-2 will get the amount and for this Consumer-2 will give up
{𝑤 − 𝑥 } units of good-2 and Consumer-1 will get the amount. The trade will continue until there are
no more mutually advantageous trade available. At that point the indifference curve of the both
consumer tangent.

From this point if anyone move to any direction, he will be better off and the other will be worse off.
Such an allocation is called Pareto Efficient Allocations.

A Pareto efficient allocation is such an allocation when it is not possible to make someone better
off without making someone else worse off.

There can be multiple Pareto efficient allocation in an Edgeworth box. The line that connects all the
Pareto efficient points is called the contract curve.

A Pareto efficient allocation can be described as an allocation where:

1. There is no way to make all the people involved better off.

2. There is no way to make some individual better off without making someone else worse off.
3. All of the gains from trade have been exhausted.
4. There are no mutually advantageous trades to be made.

The Algebra of Equilibrium

In the market the equilibrium price for both goods is –

𝑃 = 𝑃 ∗ , 𝑃∗

And the demands for good-1 of both consumers are-

𝑋 (𝑃∗ , 𝑃∗ )

𝑋 (𝑃∗ , 𝑃∗ )

© Md Khaledur Rahman
 P a g e | 18

We know,

𝐸𝑥𝑐𝑒𝑠𝑠 𝐷𝑒𝑚𝑎𝑛𝑑 = 𝐷𝑒𝑚𝑎𝑛𝑑 𝑜𝑓 𝑎 𝑔𝑜𝑜𝑑 – 𝑆𝑢𝑝𝑝𝑙𝑦 𝑜𝑓 𝑡ℎ𝑒 𝐺𝑜𝑜𝑑

So, the excess demand of the persons for a good is—

𝑒 = 𝑥 − 𝜔
𝑒 = 𝑥 − 𝜔
𝑒 = 𝑥 − 𝜔
𝑒 = 𝑥 − 𝜔

We know that the total demand of a good is equal to the total endowment of the good. Thus,

𝑋 (𝑃∗ , 𝑃 ∗ ) + 𝑋 (𝑃∗ , 𝑃∗ ) = 𝜔 + 𝜔

𝑜𝑟, [𝑋 (𝑃∗ , 𝑃∗ ) − 𝜔 ] + [𝑋 (𝑃∗ , 𝑃∗ ) − 𝜔 ] = 0

That means, in equilibrium, the sum of the excess demands of a good is equal to zero.

The total/aggregate excess demand is expressed as 𝑍 𝑓𝑜𝑟 𝑔𝑜𝑜𝑑 − 1 𝑎𝑛𝑑 𝑍 𝑓𝑜𝑟 𝑔𝑜𝑜𝑑 − 2. Thus,

𝑍 (𝑃∗ , 𝑃∗ ) = 𝑒 (𝑃∗ , 𝑃∗ ) + 𝑒 (𝑃∗ , 𝑃∗ )

And

𝑍 (𝑃∗ , 𝑃∗ ) = 𝑒 (𝑃∗ , 𝑃 ∗ ) + 𝑒 (𝑃∗ , 𝑃∗ )

Walras Law
According to the Walras’ Law, the sum of the values of aggregate excess demands is equal to zero.
That means,

𝑝 𝑧 (𝑝 , 𝑝 ) + 𝑝 𝑧 (𝑝 , 𝑝 ) ≡ 0
According to Walras’ Law, if the excess demand for one good is zero the excess demand for the other
good will also be zero.

For person-A:

𝑝 𝑥 (𝑝 , 𝑝 ) + 𝑝 𝑥 (𝑝 , 𝑝 ) ≡ 𝑝 𝜔 + 𝑝 𝜔

Or,

𝑝 [𝑥 (𝑝 , 𝑝 ) − 𝜔 ] + 𝑝 [𝑥 (𝑝 , 𝑝 ) − 𝜔 ] ≡ 0

𝑝 𝑒 (𝑝 , 𝑝 ) + 𝑝 𝑒 (𝑝 , 𝑝 ) ≡ 0

For person-B:

𝑝 𝑥 (𝑝 , 𝑝 ) + 𝑝 𝑥 (𝑝 , 𝑝 ) ≡ 𝑝 𝜔 + 𝑝 𝜔

Or,

𝑝 [𝑥 (𝑝 , 𝑝 ) − 𝜔 ] + 𝑝 [𝑥 (𝑝 , 𝑝 ) − 𝜔 ] ≡ 0

© Md Khaledur Rahman
 P a g e | 19

𝑝 𝑒 (𝑝 , 𝑝 ) + 𝑝 𝑒 (𝑝 , 𝑝 ) ≡ 0

Adding the equations for agent A and agent B together and using the definition of aggregate excess
demand, 𝑧 (𝑝 , 𝑝 ) and 𝑧 (𝑝 , 𝑝 ), we have--

𝑝 [𝑒 (𝑝 , 𝑝 ) + 𝑒 (𝑝 , 𝑝 )] + 𝑝 [𝑒 (𝑝 , 𝑝 ) + 𝑒 (𝑝 , 𝑝 )] ≡ 0
𝑝 𝑧 (𝑝 , 𝑝 ) + 𝑝 𝑧 (𝑝 , 𝑝 ) ≡ 0

Calculus of Pareto Efficiency:

In this calculation we deal with two consumers and two goods. Now we calculate how to maximize the
utility of consumer A, given consumer B’s utility.

𝑀𝑎𝑥 𝑈 (𝑥 , 𝑥 )

In this case we consider the utility of consumer B is given that means it will remain constant. Thus,

𝑈 (𝑥 , 𝑥 ) = 𝑈 − − − 𝑈𝑛𝑖𝑙𝑖𝑡𝑦 𝐶𝑜𝑛𝑠𝑡𝑟𝑎𝑖𝑛 (𝑓𝑖𝑟𝑠𝑡 𝑐𝑜𝑛𝑠𝑡𝑟𝑎𝑖𝑛)

𝑥 +𝑥 =𝑤
Resource Constrain
𝑥 +𝑥 =𝑤

We have these three constrain to solve the calculus.

Our objectives are to maximize consumer A’s utility given consumer B’s utility constant.

Now we will solve the maximization problem using Lagrange

ℒ = 𝑈 (𝑥 , 𝑥 ) + 𝜆[𝑈 − 𝑈 (𝑥 , 𝑥 )] + 𝜇 (𝑤 − 𝑥 −𝑥 ) + 𝜇 (𝑤 − 𝑥 − 𝑥 )

Here, 𝜆, 𝜇 𝑎𝑛𝑑 𝜇 𝑎𝑟𝑒 𝑡ℎ𝑒 𝑚𝑢𝑙𝑡𝑖𝑝𝑙𝑖𝑒𝑟𝑠

FOC for consumer A:

𝜕𝐿 𝜕𝑈 (𝑥 , 𝑥 )
= −𝜇 =0
𝜕𝑥 𝜕𝑥

𝜕𝐿 𝜕𝑈 (𝑥 , 𝑥 )
𝑂𝑟, = = 𝜇 −−−−−−−−−−−−− 
𝜕𝑥 𝜕𝑥

Again,

𝜕𝐿 𝜕𝑈 (𝑥 , 𝑥 )
= −𝜇 =0
𝜕𝑥 𝜕𝑥

© Md Khaledur Rahman
 P a g e | 20

𝜕𝐿 𝜕𝑈 (𝑥 , 𝑥 )
𝑂𝑟, = = 𝜇 − − − − − − − − − − − − − 
𝜕𝑥 𝜕𝑥

Dividing Equation  by Equation  ----

𝜕𝑈 (𝑥 , 𝑥 ) 𝜕𝑥 𝜇
× =
𝜕𝑥 𝜕𝑈 (𝑥 , 𝑥 ) 𝜇
𝜕𝑥 𝜇
⇒ =
𝜕𝑥 𝜇

𝜇
⇒ 𝑀𝑅𝑆 =
𝜇

FOC for consumer B:

𝜕𝐿 −𝜆 𝜕𝑈 (𝑥 , 𝑥 )
= −𝜇 =0
𝜕𝑥 𝜕𝑥

𝜕𝐿 −𝜆 𝜕𝑈 (𝑥 , 𝑥 )
𝑂𝑟, = = 𝜇 −−−−−−−−−−−−− 
𝜕𝑥 𝜕𝑥

Again,

𝜕𝐿 −𝜆 𝜕𝑈 (𝑥 , 𝑥 )
= −𝜇 =0
𝜕𝑥 𝜕𝑥

𝜕𝐿 −𝜆 𝜕𝑈 (𝑥 , 𝑥 )
𝑂𝑟, = = 𝜇 −−−−−−−−−−−−− 
𝜕𝑥 𝜕𝑥

Dividing Equation  by Equation  ----

−𝜆 𝜕𝑈 (𝑥 , 𝑥 ) 𝜕𝑥2𝐵 𝜇1
× =
𝜕𝑥 −𝜆 𝜕𝑈𝐵 (𝑥1𝐵 , 𝑥2𝐵 ) 𝜇2

𝜕𝑥 𝜇
⇒ =
𝜕𝑥 𝜇

𝜇
⇒ 𝑀𝑅𝑆 =
𝜇

We can see that the MRS of both consumer is same. That is the pareto efficient condition. Thus,

𝑀𝑅𝑆 = 𝑀𝑅𝑆

Also, it is same as price ratio. Thus,

𝑝
𝑀𝑅𝑆 =
𝑝

Because, we know higher indifference curve means higher utility and at pareto efficient point the
indifference curve tangent the budget line and the slope of the budget line is price ratio.

© Md Khaledur Rahman
 P a g e | 21

As the MRS is equal to price ratio, the MRS is also called shadow price or Pareto efficiency price.

Cost Minimization
𝑇𝑜𝑡𝑎𝑙 𝐶𝑜𝑠𝑡 = 𝐿𝑎𝑛𝑑 + 𝐿𝑎𝑏𝑜𝑟 + 𝐶𝑎𝑝𝑖𝑡𝑎𝑙 + 𝑂𝑟𝑔𝑎𝑛𝑖𝑧𝑎𝑡𝑖𝑜𝑛 𝐶𝑜𝑠𝑡

There are two types of cost: 1. Fixed cost and 2. Variable cost

Variable costs are the costs which are directly related with production. Fixed costs are the costs which
are not related with production. The producer has to face the fixed cost either he produces or do not
produce. The fixed cost is the term in short run, all costs are variable in long run.

Sunk Cost: Sunk cost are a cost that cannot be recovered. A sunk cost refers to a cost that has already
been incurred and cannot be recovered. Since it already spent and cannot be changed. [Sunk costs are
another kind of fixed costs.]

Inputs Prices
𝑥 𝑤
𝑥 𝑤

𝑀𝑖𝑛 𝐶 = 𝑤 𝑥 + 𝑤 𝑥

Such that 𝑓(𝑥 , 𝑥 ) = 𝑦

To solve this problem, we use Lagrange method. Thus,

ℒ = 𝑤 𝑥 + 𝑤 𝑥 + 𝜆[𝑦 − 𝑓(𝑥 , 𝑥 )]

FOC:

𝜕𝐿 𝜕𝑓(𝑥 , 𝑥 )
= 𝑤 −𝜆 =0
𝜕𝑥 𝜕𝑥
𝜕𝐿 𝜕𝑓(𝑥 , 𝑥 )
= 𝑤 −𝜆 =0
𝜕𝑥 𝜕𝑥
𝜕𝐿
= 𝑦 − 𝑓(𝑥 , 𝑥 ) = 0
𝜕𝜆

𝜕𝑓(𝑥 , 𝑥 )
𝑤 =𝜆 − − − −1
𝜕𝑥

𝜕𝑓(𝑥 , 𝑥 )
𝑤 =𝜆 − − − −2
𝜕𝑥
𝑦 = 𝑓(𝑥 , 𝑥 ) − − − −3

© Md Khaledur Rahman
 P a g e | 22

Dividing Equation 1 by 2
𝑤 𝜕𝑓(𝑥 , 𝑥 )/𝑑𝑥
=
𝑤 𝜕𝑓(𝑥 , 𝑥 )/𝑑𝑥

( , )/
Here is the Technical Rate of Substitution (TRS) [Output change of 𝑥 due to output
( , )/
change of 𝑥 ] and is the price ration. TRS is the slop of iso-quant curve and the price ratio is the
slop of iso cost curve. Where these two are equal the combination is least cost combination and at that
point cost is minimum.

We see another example where we apply this method to the Cobb-Douglas production function as
follows:

𝑓(𝑥 , 𝑥 ) = 𝑥 𝑥

The cost minimization problem is then,

𝑀𝑖𝑛 𝑤 𝑥 +𝑤 𝑥

Such that, 𝑥 𝑥 = 𝑦

To solve this problem, we use Lagrange method. Thus,

𝐿 = 𝑤 𝑥 + 𝑤 𝑥 + 𝜆[𝑦 − 𝑥 𝑥 ]

FOC:
𝜕𝐿
= 𝑤 − 𝜆𝑎𝑥 𝑥 =0
𝜕𝑥
𝜕𝐿
= 𝑤 − 𝜆𝑏𝑥 𝑥 =0
𝜕𝑥
𝜕𝐿
=𝑦−𝑥 𝑥 =0
𝜕𝜆

𝑤 = 𝜆𝑎𝑥 𝑥 − − − −1

𝑤 = 𝜆𝑏𝑥 𝑥 − − − −2

𝑦=𝑥 𝑥 − − − −3

Multiplying the first equation by 𝑥 and the second equation by 𝑥 to get

𝑤 = 𝜆𝑎𝑥 𝑥
𝜆𝑎𝑥 𝑥
⇒𝑤 =
𝑥

© Md Khaledur Rahman
 P a g e | 23

⇒ 𝑤 𝑥 = 𝜆𝑎𝑥 𝑥
⇒ 𝑤 𝑥 = 𝜆𝑎𝑦 𝑦=𝑥 𝑥
𝑎𝑦
⇒𝑥 =𝜆
𝑤

𝑤 = 𝜆𝑏𝑥 𝑥
𝜆𝑏𝑥 𝑥
⇒𝑤 =
𝑥
⇒ 𝑤 𝑥 = 𝜆𝑏𝑥 𝑥
⇒ 𝑤 𝑥 = 𝜆𝑏𝑦 𝑦=𝑥 𝑥
𝑏𝑦
⇒𝑥 =𝜆
𝑤

Substituting the value of 𝑥 𝑎𝑛𝑑 𝑥 in equation 3:

𝜆𝑎𝑦 𝜆𝑏𝑦
𝑦=
𝑤 𝑤
𝜆 𝑎 𝑦 𝜆 𝑎 𝑦
⇒𝑦= ×
𝑤 𝑤
λ a y .λ b y
⇒y=
w w
{one line missing}
λ a .λ b
⇒y =
w w
⇒y w w a b =𝜆

⇒λ= a b w w y

These factor demand function:

𝜆𝑎𝑦
𝑥 (𝑤 , 𝑤 , 𝑦) =
𝑤
𝑎𝑦
= a b w w y .
𝑤

=𝑎 𝑏 𝑤 𝑤 𝑦
{𝑜𝑛𝑒 𝑙𝑖𝑛𝑒 𝑚𝑖𝑠𝑠𝑖𝑛𝑔}

=𝑎 𝑏 𝑤 𝑤 𝑦
𝑎
= 𝑤 𝑤 𝑦
𝑏
Using same procedure we have,
𝜆𝑏𝑦
𝑥 (𝑤 , 𝑤 , 𝑦) =
𝑤
𝑏𝑦
= a b w w y .
𝑤

=𝑎 𝑏 𝑤 𝑤 𝑦

© Md Khaledur Rahman
 P a g e | 24

=𝑎 𝑏 𝑤 𝑤 𝑦

=𝑎 𝑏 𝑤 𝑤 𝑦
𝑎
= 𝑤 𝑤 𝑦
𝑏

The cost minimizing choice is,

𝑐(𝑤 , 𝑤 , 𝑦) = 𝑤 𝑥 (𝑤 , 𝑤 , 𝑦) + 𝑤 𝑥 (𝑤 , 𝑤 , 𝑦)
𝑎 𝑎
=𝑤 𝑤 𝑤 𝑦 +𝑤 𝑤 𝑤 𝑦
𝑏 𝑏
{𝑜𝑛𝑒 𝑙𝑖𝑛𝑒 𝑚𝑖𝑠𝑠𝑖𝑛𝑔}
𝑎 𝑎
= + 𝑤 𝑤 𝑦
𝑏 𝑏

End Here

© Md Khaledur Rahman