Commodity Pricing

Prelude
 Pricing depends on demand and supply

 Role of market structure

 Types of goods – homogeneous, heterogeneous

 Market Structure: Perfect competition, Duopoly, Oligopoly, Monopoly

etc…

2
Equilibrium
 A price quantity configuration is said to be an equilibrium if plans of
all the relevant economic agents (buyers and sellers) are realized
simultaneously

 Single commodity demand supply framework

3
Equilibrium
 At any 𝑃𝑃′ ; such that 𝑃𝑃′ > 𝑃𝑃𝑒𝑒 ; 𝑋𝑋𝑑𝑑 𝑃𝑃′ < 𝑋𝑋𝑠𝑠 𝑃𝑃′

 Quantity transacted: 𝑋𝑋𝑡𝑡 = min{𝑋𝑋𝑑𝑑 𝑃𝑃′ , 𝑋𝑋𝑠𝑠 𝑃𝑃′ }

⟹ 𝑋𝑋𝑡𝑡 𝑃𝑃′ = 𝑋𝑋𝑑𝑑 𝑃𝑃′

 Therefore at 𝑃𝑃′ buyers’ plans are realized, but not that of the sellers.

 Similarly, for any 𝑃𝑃′ < 𝑃𝑃𝑒𝑒 , sellers’ plans would be realized.

 Only at 𝑃𝑃𝑒𝑒 we have, 𝑋𝑋𝑡𝑡 𝑃𝑃𝑒𝑒 = 𝑋𝑋𝑑𝑑 𝑃𝑃𝑒𝑒 = 𝑋𝑋𝑠𝑠 𝑃𝑃𝑒𝑒

4
Equilibrium and Clearing of Market
 Let’s consider the labour market with rigid wages (may be due to
labour laws or presence of labour union); plans realised, but market
does not clear.

5
Perfectly Competitive Markets
 We define a competitive market (or perfect competition) as a market
where agents (buyers and sellers) behave competitively.

 A buyer or a seller (agent in what follows) is said to be competitive (or

alternatively, to behave competitively) if the agent assumes or believes
that the market price is given and that the agent's actions do not
influence the market price.

 Thus, the assumption of competitive behavior relates only to what

agents believe about the consequences of their actions.

 It is important to note that the assumption of competitive behavior is

independent of how many firms or consumers there are in the market
6
Perfectly Competitive Markets
 As long as the agents behave competitively, the competitive
equilibrium price can be solved for any number of buyers and sellers

 A common mix-up: assumption of competitive behavior and the

assumption that the number of sellers must be large

 Reasons for this mix-up?

1. The assumption of price-taking behavior seems more reasonable
when the number of firms is large, and each firm sells a small
amount relative to the aggregate industry sales
2. The equilibrium price solutions for some imperfectly competitive
market structures converge on (get closer to) the competitive price
when the number of firms increases
7
Perfectly Competitive Markets
 Let’s consider the inverse demand function: p (Q) = a − bQ; a, b > 0

 Given homogenous good we assume non-increasing returns to scale.

 Suppose two firms – 1 and 2

 Cost function: TCi (Q) = ci qi ; i = 1, 2; c2 ≥ c1 ≥ 0

∴ ACi (Q) = MCi (Q) = ci ∀ qi ; i = 1, 2

8
Perfectly Competitive Markets
 competitive equilibrium is a vector of quantities produced and a price
such that –
1. each firm chooses its profit-maximizing output at the given
equilibrium price, and
2. at the equilibrium price, aggregate quantity demanded equals
aggregate quantity supplied

 Formally, the triplet { p , q1 , q2 } is called the competitive equilibrium

e e e

if –
1. p , q solves
e e
i
max π i ( qi ) = p e
qi − TCi (qi ); i = 1, 2
qi

2. p e = a − b( q1e + q2e ); p e , q1e , q2e ≥ 0

9
Perfectly Competitive Markets
 Firm i treats p as a constant, the firm's profit margin defined by ( p − ci )

 Therefore, if ( p − ci ) > 0 , then the firm would produce qi = ∞

 If ( p − ci ) < 0 then qi = 0 and
 If ( p − ci ) = 0then the firm is making zero-profit at every level of
production implying that the output level is indeterminate

 Formally, the supply function would look like –

∞ if p > ci
qi = {[0, ∞) if p = ci
0 if p < ci
10
Perfectly Competitive Markets
 Now, if ( p − ci ) > 0, then qi = ∞ which violates the demand function
which ensures that quantity demanded must be finite for every price
level.
 Hence, p e ≤ c1
 However, p e < c1 ≤ c2 ⇒ q1 = q2 = 0

 If a > c 2 ≥ c1 the unique competitive equilibrium price is

, p e
= c1 and

a − c1
1. If c2 > c1 ⇒ q2 = 0, q1 =
e e

b
a − c1 e e
2. If c2 = c1 ⇒ Q = q1 + q2 = ; q1 , q2 > 0.
e e e

b
Aggregate output can be determined, but not individual output

11
Perfectly Competitive Markets
 Observe that if a < c1 (meaning that the demand is low), then neither
firm would produce.

 This model can be easily extended to any number of firms. Clearly in

equilibrium, only the firm(s) with the lowest unit cost would produce.

 competitive market structure can be imposed even if there is only one

firm. For example, if there is only one firm with a unit cost c ≥ 0 ,
a−c
then p = c and q =
e e
constitute a unique competitive
equilibrium. b

12
Perfectly Competitive Markets

13
Perfectly Competitive Markets
 Why not IRS?

F + cq if q > 0
 Let’s assume 1 firm with cost function: TC (q ) = {
F if q = 0

 Therefore, AC decreases and approaches the MC as q increases

14
Perfectly Competitive Markets
 Suppose p e = p1e ≤ c ⇒ p1e < F
q + c = ATC (q )
⇒ qe = 0
 Therefore this can’t be the equilibrium price

 Suppose p e = p2e > c ⇒ p2e > F

q + c = ATC (q )
⇒ qe = ∞

 Combining, if a>c; competitive equilibrium does not exists under IRS

15
Monopoly
 Monopoly: theory of a single seller facing competitive (price-taking)
consumers in one or several markets, over one or several periods.

 The monopolist faces a downward sloping demand curve

 The monopolist needs to devote resources to the careful study of the

demand curve facing its product – then, the monopolist can
determine either the price for the product or the quantity supplied

 After estimating the demand curve, the monopoly has to study the
market demand to determine its profit-maximizing output.

16
Monopolist’s profit maximization
 Let 𝑇𝑇𝑇𝑇(𝑄𝑄)denote the total cost function of the monopoly.

 Denoting by π(Q) the monopoly's profit level when producing Q

units of output, the monopoly chooses 𝑄𝑄𝑚𝑚 to maximize –

max 𝜋𝜋(𝑄𝑄) = 𝑇𝑇𝑇𝑇 𝑄𝑄 − 𝑇𝑇𝑇𝑇(𝑄𝑄)

𝑄𝑄

 Profit maximization (FOC) yields 𝑄𝑄𝑚𝑚 > 0 such that –

𝑑𝑑𝑑𝑑𝑑𝑑(𝑄𝑄𝑚𝑚 ) 𝑑𝑑𝑑𝑑𝐶𝐶(𝑄𝑄𝑚𝑚 )
= ⟹ 𝑀𝑀𝑅𝑅 𝑄𝑄𝑚𝑚 = 𝑀𝑀𝑀𝑀(𝑄𝑄𝑚𝑚 )
𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑

17
Monopolist’s profit maximization
 Now, 𝑀𝑀𝑅𝑅 𝑄𝑄𝑚𝑚 = 𝑀𝑀𝑀𝑀(𝑄𝑄𝑚𝑚 ) is a necessary condition, not sufficient!

 𝑀𝑀𝑅𝑅 𝑄𝑄 𝑚𝑚 = 𝑀𝑀𝑀𝑀(𝑄𝑄 𝑚𝑚 ) is only a necessary condition, meaning that if

the profit maximizing output is strictly positive, then the condition
has to be satisfied

 If the monopoly has to pay high fixed costs, it is possible that the
monopoly's profit-maximizing output level is 𝑄𝑄𝑚𝑚 = 0.

 How to ensure?

18
Monopolist’s profit maximization
 Solve for 𝑄𝑄𝑚𝑚 from, 𝑀𝑀𝑅𝑅 𝑄𝑄𝑚𝑚 = 𝑀𝑀𝑀𝑀(𝑄𝑄𝑚𝑚 )

 Plug the 𝑄𝑄𝑚𝑚 thus obtained into the total profit function to check
whether 𝜋𝜋(𝑄𝑄𝑚𝑚 ) ≥ 0.

 If 𝜋𝜋(𝑄𝑄𝑚𝑚 ) ≱ 0 for the 𝑄𝑄𝑚𝑚 obtained then monopolist chooses 𝑄𝑄𝑚𝑚 = 0

 If 𝜋𝜋(𝑄𝑄𝑚𝑚 ) ≥ 0 then the output level solved from 𝑀𝑀𝑀𝑀 𝑄𝑄𝑚𝑚 =

𝑀𝑀𝑀𝑀(𝑄𝑄 𝑚𝑚 ) is the profit-maximizing output level.

 After finding the monopoly's profit-maximizing output, the price

charged by the monopoly can be found by substituting 𝑄𝑄𝑚𝑚 into the
demand function
19
Case
 Total cost: 𝑇𝑇𝑇𝑇 𝑄𝑄 = 𝐹𝐹 + 𝑐𝑐𝑄𝑄2

 Demand curve: 𝑝𝑝 𝑄𝑄 = 𝑎𝑎 − 𝑏𝑏𝑏𝑏

 FOC: 𝑀𝑀𝑀𝑀 𝑄𝑄 𝑚𝑚 = 𝑀𝑀𝑀𝑀 𝑄𝑄 𝑚𝑚 ⟹ 𝑎𝑎 − 2𝑏𝑏𝑄𝑄 𝑚𝑚 = 2𝑐𝑐𝑄𝑄 𝑚𝑚

𝑎𝑎
 Hence, 𝑄𝑄𝑚𝑚 = 2(𝑏𝑏+𝑐𝑐)

𝑎𝑎 𝑎𝑎(𝑏𝑏+2𝑐𝑐)
 Implying, 𝑝𝑝 𝑄𝑄𝑚𝑚 = 𝑎𝑎 − 𝑏𝑏 =
2 𝑏𝑏+𝑐𝑐 2(𝑏𝑏+𝑐𝑐)

20
Case
 Therefore we have the profit function as –

2
𝑎𝑎
𝜋𝜋(𝑄𝑄 𝑚𝑚 ) ≡ 𝑇𝑇𝑇𝑇 𝑄𝑄 𝑚𝑚 − TC 𝑄𝑄 𝑚𝑚 = 𝑝𝑝𝑚𝑚 𝑄𝑄𝑚𝑚 − 𝐹𝐹 − 𝑐𝑐(𝑄𝑄 𝑚𝑚 )2 = − 𝐹𝐹
4(𝑏𝑏 + 𝑐𝑐)

 So, we can formally write down the monopolist’s profit

maximization output as –

𝑎𝑎 𝑎𝑎2
𝑄𝑄 𝑚𝑚 = � 2(𝑏𝑏 + 𝑐𝑐) 𝑖𝑖𝑖𝑖 𝐹𝐹 ≤
4(𝑏𝑏 + 𝑐𝑐)
0 𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜

21
Graphical Illustration ( 𝑄𝑄 𝑚𝑚 >0 )
 Case where the demand is high enough (or the fixed cost is low
enough) so that the monopolist would produce 𝑄𝑄𝑚𝑚 > 0

22
Graphical Illustration ( 𝑄𝑄 𝑚𝑚 =0 )
 Case where the demand is so low (or the fixed cost is high enough)
so that the monopolist would produce 𝑄𝑄𝑚𝑚 = 0

23
Discriminating Monopoly
 The analysis so far has focused, on monopolies charging a single, uniform
price to all customers.

 A monopolist can, however, increase its profit by charging different

prices to consumers with different characteristics. That is, a firm may be
able to differentiate among consumers according to tastes, income, age,
and location in order to charge consumers with different characteristics
different prices.

 In order to be able to charge consumers different prices, a monopolist

must possess the means for making arbitrage (buying low for the purpose
of reselling at a high price) impossible. Price discrimination is impossible
when those consumers who are able to purchase at a low price can make a
profit by reselling the product to the consumers who buy at high prices.

24
Discriminating Monopoly
 Monopolies resort to various marketing techniques to prevent arbitrage
from taking place (market segmentation) –

1. Charge different prices at different locations (in order for price

discrimination to be sustained, the markets should be isolated by
geography, by prohibitive taxes (such as tariffs), or by prohibitive
transportation costs).

2. Monopolies that provide services (such as transportation companies,

restaurants, and places of entertainment) charge senior citizens lower
prices than they charge younger consumers – senior citizen ID or say,
student discounts

3. Book publishers manage to charge institutions higher prices than they

charge individuals by selling hardcovers to institutions and softcovers to
individuals 25
Discriminating Monopoly
 Let’s assume that arbitrage cannot take place and a monopolist is selling in
two different markets

 Question: How does a monopolist determine the output level (and, hence,
the price) in each market?

 The monopolist chooses the output levels sold in each market, 𝑞𝑞1𝑚𝑚 and
𝑞𝑞2𝑚𝑚 , that solve –

max 𝜋𝜋(𝑞𝑞1 , 𝑞𝑞2 ) = 𝑇𝑇𝑇𝑇1 𝑞𝑞1 + 𝑇𝑇𝑇𝑇2 𝑞𝑞2 − 𝑇𝑇𝑇𝑇(𝑞𝑞1 , +𝑞𝑞2 )

𝑞𝑞1 ,𝑞𝑞2

 FOCs: 𝑀𝑀𝑅𝑅𝑖𝑖 𝑞𝑞𝑖𝑖𝑚𝑚 = 𝑀𝑀𝑀𝑀 𝑞𝑞1𝑚𝑚 + 𝑞𝑞2𝑚𝑚 ∀ 𝑖𝑖 = 1, 2

26
Discriminating Monopoly
 If the monopolist chooses 𝑞𝑞1𝑚𝑚 and 𝑞𝑞2𝑚𝑚 such that 𝑀𝑀𝑀𝑀1 𝑞𝑞1𝑚𝑚 > 𝑀𝑀𝑀𝑀2 𝑞𝑞2𝑚𝑚 ,
then, it is clear that the monopoly should transfer one unit from market 2
to market 1.

 In this case the reduction in revenue in market 2 is smaller than the

increase in revenue in market 1.

 To solve for the profit-maximizing output levels 𝑞𝑞1𝑚𝑚 and 𝑞𝑞2𝑚𝑚 , we need to
solve two equations (FOCs) with the two variables.

 Graphically – 3 steps

1 1
 Prices charged 𝑝𝑝1𝑚𝑚 and 𝑝𝑝2𝑚𝑚 such that 𝑝𝑝1𝑚𝑚 1 + = 𝑝𝑝2𝑚𝑚 1 +
𝜀𝜀1 𝜀𝜀2

27
Discriminating Monopoly

28
Discriminating Monopoly
 Implication: A discriminating monopoly selling a strictly positive
amount in each market will charge a higher price at the market with
the less elastic demand

 If 𝜀𝜀2 > 𝜀𝜀1 (or |𝜀𝜀2 |<|𝜀𝜀1 |) then 𝑝𝑝2𝑚𝑚 > 𝑝𝑝1𝑚𝑚

29
Cartel and Multi-plant Monopoly
 The cartel and the multi-plant monopoly are forms of organizations and
contractual agreements among plants, firms, or countries

 Cartel: cartel is an organization that contracts with the plants on how much
each would produce and hence on what would be the price

 Examples: OECD (assuming each oil producing country is a plant), IATA,

Bar Associations

 The multiplant monopoly is very similar to the cartel, except that all the
plants are put under a single ownership.
 Multiplant monopoly occurs when several firms in the industry merge
together into a single firm (horizontal merger), or when a monopoly firm
opens several plants producing the same product.

30
Cartel and Multi-plant Monopoly
 Thus, unlike the cartel, the multiplant monopoly has the power to decide
whether to shut down some of its plants (or whether to open several
more).

 A cartel generally does not shut down plants or countries for the simple
legal reason that the cartel does not own the plants, and no plant would
join the cartel knowing that it could be shut down.

 We assume a linear aggregate demand given by 𝑝𝑝 = 𝑎𝑎 − 𝑏𝑏𝑏𝑏.

 We assume that there are N plants, indexed by i, (i = 1, 2,..., N).

 Each plant produces 𝑞𝑞𝑖𝑖

31
Cartel and Multi-plant Monopoly
 Let’s assume that each plant has the technology summarized by the total
cost function given by –

𝑇𝑇𝑇𝑇𝑖𝑖 𝑞𝑞𝑖𝑖 = 𝐹𝐹 + 𝑐𝑐(𝑞𝑞𝑖𝑖 )2 ; 𝐹𝐹, 𝑐𝑐 > 0

 Thus, we assume that all plants. have identical cost functions, and that each
plant has a fixed (output independent) cost of F.

 The plant's average and marginal-cost functions are given by –

𝐹𝐹
𝐴𝐴𝐴𝐴𝐶𝐶𝑖𝑖 𝑞𝑞𝑖𝑖 = + 𝑐𝑐𝑞𝑞𝑖𝑖
𝑞𝑞𝑖𝑖
𝑀𝑀𝐶𝐶𝑖𝑖 𝑞𝑞𝑖𝑖 = 2𝑐𝑐𝑞𝑞𝑖𝑖

32
Cartel
 The cartel organizes all the N plants by directing each plant to produce a
certain amount.

 The objective is to maximize the Sum of the profits of all the N plants.

 Let’s denote 𝜋𝜋𝑖𝑖 𝑞𝑞𝑖𝑖 be the profit of the ith plant and 𝑄𝑄 = ∑𝑁𝑁
𝑖𝑖=1 𝑞𝑞𝑖𝑖

 Objective of the cartel: Choose 𝑞𝑞1 , 𝑞𝑞2 , ……, 𝑞𝑞𝑁𝑁 such that –
𝑁𝑁

max 𝜋𝜋 𝑞𝑞1 , 𝑞𝑞2 , ……, 𝑞𝑞𝑁𝑁 ≡ � 𝜋𝜋𝑖𝑖 𝑞𝑞𝑖𝑖

𝑞𝑞1 , 𝑞𝑞2 ,
……, 𝑞𝑞𝑁𝑁
𝑖𝑖=1
𝑁𝑁 𝑁𝑁 𝑁𝑁

= [𝑎𝑎 − 𝑏𝑏 � 𝑞𝑞𝑖𝑖 ] � 𝑞𝑞𝑖𝑖 − � 𝑇𝑇𝑇𝑇𝑖𝑖 𝑞𝑞𝑖𝑖

𝑖𝑖=1 𝑖𝑖=1 𝑖𝑖=1

33
Cartel
𝜕𝜕𝜕𝜕
 N FOCs: = 0 = 𝑎𝑎 − 2𝑏𝑏 ∑𝑁𝑁
𝑖𝑖=1 𝑞𝑞𝑖𝑖 − 𝑀𝑀𝑀𝑀𝑗𝑗 𝑞𝑞𝑗𝑗
𝜕𝜕𝑞𝑞𝑗𝑗
= 𝑀𝑀𝑀𝑀 𝑄𝑄 − 𝑀𝑀𝑀𝑀𝑗𝑗 𝑞𝑞𝑗𝑗 ; 𝑗𝑗 = 1, 2, … , 𝑁𝑁

 Implication: The cartel's profit-maximizing output produced by each plant is

found by equating the marginal revenue function (derived from the market demand
curve, evaluated at the aggregate cartel-output level) to the marginal-cost function
of each plant.

 Since all plants have identical cost functions, we search for a symmetric
equilibrium where the cartel directs each plant to produce the same
𝑎𝑎
output level. Hence, 𝑞𝑞𝑗𝑗 = 𝑞𝑞 ∀ 𝑗𝑗, and we have 𝑞𝑞 = 2(𝑏𝑏𝑏𝑏+𝑐𝑐) such that
𝑁𝑁𝑎𝑎 𝑎𝑎(𝑏𝑏𝑏𝑏 + 2𝑐𝑐)
𝑄𝑄 = 𝑎𝑎𝑎𝑎𝑎𝑎 𝑝𝑝 = 𝑎𝑎 − 𝑏𝑏𝑏𝑏 =
2(𝑏𝑏𝑏𝑏 + 𝑐𝑐) 2(𝑏𝑏𝑏𝑏 + 𝑐𝑐)

34
Cartel
 Note that for 𝑁𝑁 = 1 the cartel's output and price coincide with the pure
monopoly levels.

 It can be easily verified that as the number of firms in the cartel increases
(N increases), both the output level of each firm and the market price fall
(q and p decrease).

 Hence, the total revenue and profit of each firm must fall with an increase
in the number of cartel members.

 For this reason, many professional organizations, such as those of lawyers

and accountants, impose restrictions on new candidates who wish to
practice in their profession.

35
Multi-plant Monopoly
 The multiplant monopoly is very similar to the cartel, except that it has the
authority (ownership) to shut down some plants, thereby "saving" variable
and fixed costs associated with maintaining the plant.

 Thus, if we suppose that the multiplant monopoly can choose the number
of plants, that is, N is a choice variable by the multiplant monopoly owner,
then the question is: What is the profit maximizing number of plants
operated by the multiplant monopoly?

 Given that the multiplant monopoly can add or discard plants, the
monopoly would seek to adjust the number of plants to minimize the cost
per unit of production

36
Multi-plant Monopoly
 In other words, the multiplant monopoly will adjust the number of plants
to minimize 𝐴𝐴𝐴𝐴𝐴𝐴𝑖𝑖 𝑞𝑞𝑖𝑖 for every plant in operation.

 Let N be continuous (quite an assumption!!)

 Like the cartel, the multiplant monopoly would equate 𝑀𝑀𝑀𝑀 𝑄𝑄 − 𝑀𝑀𝑀𝑀𝑖𝑖 𝑞𝑞𝑖𝑖
𝑎𝑎
for every operating plant, yielding, 𝑞𝑞𝑖𝑖 =
2(𝑏𝑏𝑏𝑏+𝑐𝑐)

 In addition, the monopoly will adjust N so that each operating plant would
operate at minimum 𝐴𝐴𝐴𝐴𝐴𝐴𝑖𝑖 𝑞𝑞𝑖𝑖 where min 𝐴𝐴𝐴𝐴𝐴𝐴𝑖𝑖 𝑞𝑞𝑖𝑖 = 𝐹𝐹⁄𝑐𝑐

𝑎𝑎 𝑎𝑎 𝑐𝑐 𝑐𝑐
 Hence, 𝐹𝐹⁄
𝑐𝑐 = yields 𝑁𝑁 𝑚𝑚 = −
2(𝑏𝑏𝑏𝑏+𝑐𝑐) 2𝑏𝑏 𝐹𝐹 𝑏𝑏

37
Multi-plant Monopoly
 Thus, the multiplant monopoly's profit-maximizing number of plants
increases with an increase in the demand parameter a, and decreases with
the fixed cost parameter of each plant F.

38
Heterogeneous Commodities
 Consider an industry with fairly large number of firms (electronic
goods, banks and financial institutions, newspapers, medicines etc..)

 Theories suggest they will earn zero profit (perfect competition,

Bertrand oligopoly) or near zero profits (competitive selection) in the
long-run

 Evidence: all of them earn fairly handsome profits in the long-run,

which can’t exit.

 One of the main source of such profit despite of large number of

players is differentiated products.

 Basic logic: different tastes, different varieties and hence become a

price maker in at least one variety and have your own niche.

39
Observations
1. Most industries produce a large number of similar but not
identical products.

2. Only a small subset of all possible varieties of differentiated

products are actually produced. For example, most products are
not available in all colors.

3. Most industries producing differentiated products are

concentrated, in the sense that it is typical to have two to five
firms in an industry.

4. Consumers purchase a small subset of the available product

varieties.

40
Types and Modeling
 Horizontal: Different types of consumers prefer different types of
products (preferring Coke over Pepsi) – consumers have brand
preferences. Includes location or address models (buyers buy only
preferred products) and non-address Chamberlinian models.

 Vertical: Unanimous ranking of products by the consumers (say, Mac is

better than any other laptop brands) – consumers can rank quality and
buy the best if prices are equal

 Non-address models: all consumers derive utility from consuming a

variety of products – no brand loyalty

 Characteristics: Consumers prefer not the products(s) but embedded

characteristics. (1) more characteristics better (2) some characters are
preferred (3) varieties

41
Simple (non-Address) Models
 Consider a two-firm industry producing two differentiated products
indexed by i = 1, 2.

 To simplify the exposition, we assume that production is costless.

 Let the inverse demand functions to be:

p1 = α − βq1 − γq2 ; p2 = α − γq1 − βq2 ; β > 0; β 2 > γ 2

 Thus, we assume that that there is a fixed number of two brands and
that each is produced by a different firm facing an inverse demand
curve

 The last condition implies own price elasticity larger than cross price
elasticity.

42
Simple (non-Address) Models
 Direct demand functions:

q1 = a − bp1 + cp2 ; q2 = a + cp1 − bp2

 Find values of a, b and c.

α (β − γ )
a= 2
β −γ 2
β
b= 2 >0
β −γ 2

γ
c= 2
β −γ 2

43
Measure of Differentiation
 Brand’s measure of differentiation is given by

γ2
δ= 2
β

 Brands are said to be highly differentiated when the change in one

brand’s price has negligible effect on the demand of the other brand,
that is, when γ → 0 and therefore, δ → 0.

 Brands are quite homogeneous when the cross and own price effects
are almost the same. That is, price of both the brands affect each others
prices significantly. Technically γ 2 → β 2 and hence, δ → 1.

44
Zones of Differentiation

45
Quantity Game
 Cournot market structure – firms choose quantity produced as
strategies

 we look for a Nash equilibrium in firms' output levels

 zero production costs

 inverse demand functions: p1 = α − βq1 − γq2

p2 = α − γq1 − β q2
β > 0; β 2 > γ 2

46
Quantity Game
 Each firm i takes q j as given and chooses qi to

max qi π i (qi , q j ) = (α − βqi − γq j )qi ; i = 1, 2; i ≠ j

∂π i
 FOCs: (qi , q j ) = (α − 2 βqi − γq j ) = 0
∂qi

 Hence, the best response functions can be written as –

α − γq j
qi = Ri (q j ) =
2β

47
Quantity Game
 Depiction of the best response functions:

48
Quantity Game
α γ
 the best response functions are: R1 (q2 ) = − q2
2β 2β
α γ
R2 (q1 ) = − q1
2β 2β

 Note that as γ → β (the products are more homogeneous), the best-

response function becomes steeper, thereby making the profit-
maximizing output level of firm i more sensitive to changes in the
output level of firm j (due to stiffer competition).

 When γ → 0 (the products are more differentiated), the best-response

function becomes constant (zero sloped)

49
Quantity Game
 Solving the best-response functions we have –

α αβ α β 2
q =
c
; pi =
c
; πi =
c

2β + γ 2β + γ (2β + γ ) 2
i

 Note that as γ increases (the products are leas differentiated), the

individual and aggregate quantity produced, the prices, and the profits
all decline

 In a Cournot game with differentiated products, the profits of

firms increase when the products become more differentiated

50
Quantity Game
 This result can explain why firms tend to spend large sums of money
to advertise their brands

 Firms would like the consumers to believe that the brands are highly
differentiated from the competing brands for the purpose of increasing
their profits.

 In other words, differentiation increases the monopoly power of

brand-producing firms

51
Price Game
 Bertrand market structure, firms choose prices as their strategies

 we look for a Nash equilibrium in firms' price levels

 zero production costs

 direct demand functions: q1 = a − bp1 + cp2 ; q2 = a + cp1 − bp2

α (β − γ ) β γ
 Remember: a= 2 ;b= 2 > 0; c = 2
β −γ 2
β −γ 2
β −γ 2

52
Price Game
 Each firm i takes p j as given and chooses pi to

max pi π i ( pi , p j ) = (a − bpi + cp j ) pi ; i = 1, 2; i ≠ j

∂π i
 FOCs: ( pi , p j ) = (a − 2bpi + cp j ) = 0
∂pi

 Hence, the best response functions can be written as –

a + cp j
pi = Ri ( p j ) =
2b

53
Price Game
 Depiction of the best response functions:

54
Price Game
 Note that there is something different between the best response
functions obtained in the Bertrand game vis-à-vis the Cournot game.

 In price games, the best-response functions are upward sloping,

implying that, if one firm raises its price, the other would respond by
raising its price as well

 Definitions –

1. Players' strategies are said to be strategic substitutes if the best-

response functions are downward sloping
2. Players' strategies are said to be strategic complements if the best-
response functions are upward sloping

55
Price Game
 Hence, in a quantity game the quantities are strategic substitutes,
whereas in a price game prices are strategic complements.

 Quite intuitive: both are rational players maximizing profit

 Now, solving the best-response functions we get –

a α (β − γ )
p =
b
=
2b − c 2β − γ
i

ab αβ
qi =
b
=
2b − c (2β − γ )( β + γ )
a 2
b α 2
β (β − γ )
πi =
b
=
(2b − c) 2
(2β − γ ) ( β + γ )
2
56
Price Game
 The profit levels decline when the products become less differentiated
(γ increases).

 In the limit, when γ = β, the products become homogeneous, and the

profits drop to zero

 In a Bertrand game with differentiated products, the profits of

firms increase when the products become more differentiated.

 As with the Cournot case, product differentiation increases the

monopoly power of brand-producing firms by loosening up price
competition among the brand-producing firms

57
Cournot versus Bertrand
 Which market structure, a Cournot or a Bertrand, would yield a
higher market price?

 How would changing the degree of product differentiation affect the

relative difference between the two market structure outcomes?

 Formal comparison yields –

αβ α (β − γ ) α
p −p =
c b
− = > 0  [ β > γ ]
2 2
i
2β + γ (2β − γ )
i
β 2
4 2 −1
γ

58
Cournot versus Bertrand
 Thus, in a differentiated products industry –

1. The market price under Cournot is higher than it is under Bertrand.

Formally, p > p
c b
i i

2. The more differentiated the products are, the larger the difference
between the Cournot and Bertrand prices.
Formally, ∂ ( pic − pib )
>0
∂γ

3. This difference in prices is zero when the products become

independent.
Formally, limγ →0 [ pic − pib ] = 0
59
Cournot versus Bertrand
 Under Cournot market structure each firm expects the other firm to
hold its output level constant.

 Hence, each firm would maintain a low output level since it is aware
that a unilateral output expansion would result in a drop in the market
price.

 In contrast, under the Bertrand market structure each firm assumes

that the rival firm holds its price constant, hence output expansion will
not result in a price reduction.

 Therefore, more output is produced under the Bertrand market

structure than under the Cournot market structure.

60
Sequential Game
 So far we have analyzed industries where firms strategically choose
their output/price levels.
 Both the games were static in the sense that players simultaneously
choose their price/quantity produced.
 Let’s we assume that the firms move in sequence.
 Best known example is the Stackelberg game which is a sequential
version of the Cournot game (in a two-firm, sequential moves game,
firm 1 will choose its output level before firm 2 does. Then, firm 2,
after observing the output level chosen by firm 1, will choose its
output level, and only then will output be sold and profits collected by
the two firms).
 This type of market structure is often referred to as Leader-Follower
structure

61
Sequential Game
 However we try to solve for a sequential game in prices (for the
original Stackelberg game see section 6.2)

 We analyze a two-stage game, where firm 1 (the leader) chooses the

price in the first stage .
 The price chosen in the first stage is irreversible and cannot be
adjusted in the second stage.
 In the second stage, only firm 2 (the follower) chooses its price after
observing the price chosen by firm 1 in the first stage.
 Here, the game ends after the second stage, and each firm collects its
profit.

62
Sequential Game
 Our main questions are –

1. Is there any advantage for moving in the first stage rather than the
second? (One related and important question is who moves first?
That we will analyze later with an established firm and a new
entrant)

2. How would the outcomes (price, profit, quantity) compare to the

static outcomes?

 Approach – backward induction (SPNE)

63
Sequential Game
 Direct demand functions: q1 = 168 − 2 p1 + p2
q2 = 168 + p1 − 2 p2

 For these demand functions the static Bertrand outcomes are –

pib = 56; qib = 112; π ib = 6272

 In the first period, firm 1 takes firm 2's best-response function as

given, and chooses p1 that solves
168 + p1
max p1 π 1 ( p1 , R2 ( p1 )) = (168 − 2 p1 + ) p1
4

64
Sequential Game
 FOC: ∂π 1 7
= 210 − p1 = 0
∂p1 2
⇒ p1s = 60
⇒ p2s = 57
⇒ q1s = 105; q2s = 114
⇒ π 1s = 6300; π 2s = 6498

 Therefore, π 1s > π 1b ; π 2s > π 2b

65
Sequential Game
 Under a sequential-moves price game (or more generally, under any
game where actions are strategically complements):

1. Both firms collect a higher profit under a sequential-moves game

than under the single-period Bertrand game. Formally, π >π
i
s b
i

2. The firm that sets its price first (the leader) makes a lower profit
than the firm that sets its price second (the follower).

3. Compared to the Bertrand profit levels, the increase in profit to the

first mover (the leader) is smaller than the increase in profit to the
second mover (the follower). Formally, π 1s − π 1b < π 2s − π 2b

66
Sequential Game
 Thus, first to move is not always an advantage.

 Here, each firm would want the other firm to make the first move.

 The intuition behind this result is as follows.

 When firm 1 sets its price in period 1, it calculates that firm 2 will
slightly undercut p1 in order to obtain a larger market share than
firm 1.
 This calculation puts pressure on firm 1 to maintain a high price to
avoid having firm 2 set a very low market price.
 Hence, both firms set prices above the static Bertrand price levels.
 Now, firm 1 always makes a lower profit than firm 2, since firm 2
slightly undercuts firm 1 and captures a larger market share.
67
Sequential Game
 Finally, note that we could have predicted that the profit of firm 1 will
increase beyond the static Bertrand profit level even without resorting
to the precise calculations.

 Using a revealed profitability argument, we can see clearly that firm 1

can always set p1 = p and make the same profit as under the static
b
1
Bertrand game.
 However, given that firm 1 chooses a different price, its profit can only
increase.
 The profit of firm 2 (the follower's) is higher under the sequential-
moves price game than its profit under the static Bertrand game.
 In contrast, under the sequential-moves quantity game the followers'
profit is lower than it is under the static Cournot game.

68
Address Model (Concepts)
 Each person prefers a particular quality. Suppose you prefer a particular
shade of blue emulsion paint. Anything else (even other shades of
blue) you don’t prefer. Now if such shade is not available in the market
then you will have to deviate (may be slightly). In doing so you incur
some cost.

 Location:
1. Physical location (consumer checks the price in all the stores and
buys from that in which price plus transportation cost is minimum).
2. Distance between brand characteristics from that of the “ideal”
brand. Consumer disutility arises from buying this less-than-ideal
brand.
3. Brands are not uniformly ranked by all consumers.
4. Given same prices, consumers will buy their preferred brands.

69
Linear City Model (Hotelling)
 Consider a linear street of length L > 0

 Consumers residing on that street are uniformly distributed ⇒ n = L

 Each consumer is indexed x ∈ [0, L] ⇒ consumer x is located at a

distance x from the origin.

70
Price Game with Fixed Location
 Two firms, homogeneous product, location of product sold is different,
zero production cost.

 Assume: Firm A located a distance a from the origin and firm B located
at a distance b from L.

71
Price Game with Fixed Location
 Each consumer buys only one product.

 Unit transportation cost is t.

 Utility function of a consumer located at x is given by –

U x = − pA − t | x − a | if he buys from A
U x = − p B − t | x − ( L − b) | if he buys from B

 Consumer at x is indifferent between buying from A and B.

 Formally if a < x < L−b then,

− p A − t ( x − a) = − pB − t ( L − b − x)

72
Price Game with Fixed Location
pB − p A L − b + a
 Or, x= +
2t 2
 Anyone left of this point will buy from A

 On the right will buy from B

pB − p A L − b + a
 Hence, x= + gives the demand function faced
by A 2t 2

p A − pB L + b − a
 Similarly for firm B we get L−x = +
2t 2
 Assuming Bertrand competition find NE price strategies.

73
Price Game with Fixed Location
pB − p A L − b + a
 Firm A: π A = [ + ] pA
2t 2
pB − 2 p A L − b + a
 FOC: + =0
2t 2
p A − 2 pB L + b − a
 Similarly for firm B: + =0
2t 2
t (3L − b + a ) N t (3L + b − a )
 NE prices: p = ; pB =
N
A
3 3
3L − b + a
 Equilibrium market share of A: x A =
N

6
t (3L − b + a ) 2
 Equilibrium profit of A: π N
A =p x =
N
A
N
A >0
18

74
Price Game with Fixed Location
 If a = b then equal market share.

t (3 L − b + a ) 2
 The profit of firm A: π A = p AN x AN = >0
N

18

 This shows that the profit of each brand-producing firm

increases with the distance between the firms.

 firms reach higher profit levels when the brands they

produce are more differentiated

 What happens if a + b = L ?

75
Price Game with Fixed Location
 If both firms are located at the same point ( a + b = L) ,
meaning that the products are homogeneous), then the
unique equilibrium is p A = p B = 0

 A unique equilibrium exists and is described by A

N N N N
p ; p B ; x A ; x B
if and only if the two firms are not too close to each other

 Formally if and only if 2

a − b   4 L(a + 2b) 

L+  ≥ 
 3   3 
and
b − a   4 L(a + 2b) 
2

L+  ≥ 
 3   3 

76
Price Game with Fixed Location
 We consider the simple case where firms are located at
equal distances along the edges.

 That is, assume that ( a = b; a < L / 2)

 we need to show that the equilibrium exists if and only if

L2 ≥ 4 La or a ≤ L / 4

 When a = b, the distance between the two firms is (L - 2a)

 Hence, if equilibrium exists, we must have p A = p B = tL

77
Price Game with Fixed Location
 The profit level of firm A as a function of its own price p A
and a given B's price p B = tL for the case of a = b

78
Price Game with Fixed Location
 Region I: p A < tL − t ( L − 2a )
In this zone p A is very low, so that even the consumer
located at the same point where firm B is located would
purchase from firm A.
Thus, firm A has the entire market, and its profit is given
by π A = p A L = [tL − t ( L − 2a ) − ε ]L

 Region II: Here, both firms sell a strictly positive amount,

so the profit of firm A as a function of p A is given by
pB − p A L − b + a
πA =[ + ] pA
2t 2
( pA )2
Substituting p B = tL we have π A = p A L −
Maximization wrt p A yields π A = tL / 2
2 2t
79
Price Game with Fixed Location
 Region III: Exact opposite of Region I.
In this zone p A is very high, all consumers purchase from
firm B.

 Now, for a given p B = tL; π A has two local maxima.

 But in Region I, Firm A has the entire market.

N N N N
 Thus to have an eqm defined by A
p ; p B ; x A ; x B; the
globally profit-maximizing price for firm A would lie in
Region II

80
Price Game with Fixed Location
 Hence, we must have profit in region II higher than profit
in region I

 Formally, [tL − t ( L − 2a )]L ≤ tL / 2 ⇒ a ≤ L / 4

2

 When the two firms are located too closely, they start
undercutting each other's prices, resulting in a process of
price cuts that does not converge to an equilibrium.

 Thus, for an equilibrium to exist, the firms cannot be too

closely located

81
Price Game with Fixed Location
 So far, we have assumed that the location of the firms is
fixed

 What if firms can choose price and location both?

 Unfortunately, we now show that there is no solution for

this two-dimensional strategy game.

 To show that, we ask what would firm A do if, given the

price and location of its opponent, it would be allowed to
relocate.

82
Price Game with Fixed Location
 Let’s reconsider the profit function
t (3 L − b + a ) 2
π AN = p AN x AN = >0
18

 Then for any locations a and b, can firm A increase its

profit by moving toward firm B (obviously, to gain a higher
market share) or
∂π A
>0
∂a

 This case, where firms tend to move toward the center, is

called in the literature the principle of minimum
differentiation since by moving toward the center the
firms produce less-differentiated products.
83
Price Game with Fixed Location
 However, we have already seen that if firm A gets too close
to firm B, an equilibrium will not exist.

 Also, if firm A locates at the same point where firm B

locates, its profit will drop to zero

 implying that it is better off to move back to the left

 In the Hotelling linear-city game, there is no

equilibrium for the game where firms use both prices
and location as strategies.

84
Reference
 Oz Shy (1995). Industrial Organization. MIT Press. Chapters 2, 4 - 7.

85