Managerial Decision Making

Recitation class

Luiss University

Instructor: Luca Vitali [email protected]

Teaching Assistant: Diletta Topazio [email protected]

1) A Cournot Game with asymmetric costs
Consider a duopoly where the inverse demand is 𝑝 = 𝑝 𝑞 . The unit cost
faced by the two firms is, respectively, 𝑐! and 𝑐" .Firms choose individually a
quantity to produce (denoted 𝑞! and 𝑞" ), thereafter the selling price is set to
𝑝 𝑞! + 𝑞" .
a) Compute the firm’s payoff and explain why the two firms are in strategic
interaction

In the general formulation:

𝜋! = 𝑝 𝑞! + 𝑞" − 𝑐! 𝑞!
𝜋" = 𝑝 𝑞! + 𝑞" − 𝑐" 𝑞"

The firms are in strategic interaction, as for example, firm 1 ’s profit depend
not only on 𝑞! but also on 𝑞" .
b) Consider that 𝑝 𝑞 = 1 − 𝑞 and write the profit functions
𝜋! = 1 − 𝑞! − 𝑞" − 𝑐! 𝑞!
𝜋" = 1 − 𝑞! − 𝑞" − 𝑐" 𝑞"

c) Compute reaction functions for each firm and the Nash equilibrium
𝜕𝜋!
= 1 − 2𝑞! − 𝑞" − 𝑐! = 0
𝜕𝑞!
𝜕𝜋"
= 1 − 𝑞! − 2𝑞" − 𝑐" = 0
𝜕𝑞"
and then, we find the reaction function of firm 𝑖 as:
1
𝑞! = 1 − 𝑞" − 𝑐!
2
1
𝑞" = 1 − 𝑞! − 𝑐"
2
By substitution we get:
1 1
𝑞! = 1 − 1 − 𝑞! − 𝑐" − 𝑐!
2 2
1 1 1 1
𝑞! = 1 − + 𝑞! + 𝑐" − 𝑐!
2 2 2 2
1 1 1 1
𝑞! = + 𝑞! + 𝑐" − 𝑐!
4 4 4 2
4 1 1 1
𝑞! = + 𝑐" − 𝑐!
3 4 4 2
∗
1 2
𝑞! = 1 + 𝑐" − 𝑐!
3 3
and by symmetry:
1 2
𝑞"∗ = 1 + 𝑐! − 𝑐"
3 3
d) Check that 𝑞!∗ > 𝑞"∗ and interpret.
We compute the difference between 𝑞!∗ and 𝑞"∗

1 1 2 1 1 1
𝑞!∗ − 𝑞"∗ = + 𝑐" − 𝑐! − − 𝑐! + 𝑐" = 𝑐" − 𝑐!
3 3 3 3 3 3

If 𝑐" > 𝑐! then 𝑞!∗ > 𝑞"∗ , since 𝑐" > 𝑐! means that firm 1 is more efficient than
firm 2.
The Cournot market use more efficient firms.
e) Check the conditions under which 𝑞"∗ = 0
∗ " ! !
𝑞" = 0 means that 𝑐" = + 𝑐! , i.e.
$ $ $
1
𝑐" = 1 + 𝑐!
2
So, when the difference in costs is quite large and firm 2 is so much less
efficient than firm 1, then firm 2 does not enter the market.
 !
f) Describe the market under the condition 𝑐" ≥ 1 + 𝑐! . Find and interpret
"
the behavior of firm 1.
∗ !
In this case, 𝑞! = 1 − 𝑐!
"
In order to show that this corresponds to a monopoly behavior we compute
the corresponding price-cost margin. Notice that the price is
∗
1 1 1
𝑝 = 1 − 𝑞 = 1 − − 𝑐! = 1 + 𝑐!
2 2 2
the monopolist’s optimal markup of price over marginal cost, expressed as a
percentage of the price is
1 1
𝑝 − 𝑐! 2 1 + 𝑐! − 𝑐! 2 1 − 𝑐!
∗
1 − 𝑐!
= = =
𝑝 ∗ 1 1 1 + 𝑐!
2 1 + 𝑐! 2 1 + 𝑐!
This is exactly the opposite of the inverse of the elasticity of demand of the
market computed when the quantity is 𝑞! . Indeed,
𝑞 =1−𝑝
and price elasticity of demand is
𝛿𝑞 𝑝 1 + 𝑐!
𝜀%/' = − =
𝛿𝑝 𝑞 1 − 𝑐!
 Bertrand competition
Two firms produce a homogeneous good.
Cost function for both firms is 𝐶 𝑞! = 𝑐𝑞! , where 𝑞! is the quantity
sold by firm 𝑖.
Firms compete by simultaneously and independently setting the
price, ready to sell any demanded quantity.
No capacity constraints.
Demand function:
𝑄 =𝛼−𝑝 with 𝑝 = 𝑚𝑖𝑛 𝑝" , 𝑝# and 𝛼 > 𝑐

In case they set the same price then consumers’ demand is split
evenly between them.
 Game in Normal Form
Set of players: firm 1 and firm 2
Each firm can choose among non-negative prices, 𝑝! ≥ 0
Set of strategies: 𝑆! = [0, ∞) for 𝑖 = 1,2
Payoff to firm 1:

a) 𝑖𝑓 𝑝" < 𝑝# → 𝜋" = 𝑝" − 𝑐 𝛼 − 𝑝"

b) 𝑖𝑓 𝑝" > 𝑝# → 𝜋" = 0

"
c) 𝑖𝑓 𝑝" = 𝑝# → 𝜋" = 𝑝" − 𝑐
#
𝛼 − 𝑝"
1) Neither firm can be pricing below 𝑀𝐶 = 𝑐 in equilibrium
because in this case at least one of the firms earns a negative
profit, and each firm can guarantee itself a profit of 0 by pricing
at 𝑐.

2) 𝑝" > 𝑝# ≥ 𝑐 cannot be an equilibrium. If 𝑝# > 𝑐, then firm 1 can

raise its profit (from 0) by changing its price to be between 𝑐 and
𝑝# . Further, if 𝑝# = 𝑐, then firm 2 can increase its profit by raising
its price while still undercutting the price charged by firm 1.

3) 𝑝" = 𝑝# > 𝑐 cannot be an equilibrium because then each firm

gets half of the market demand but could grab the whole
demanded quantity by dropping its price by a tiny amount.
 Nash Equilibria in pure strategies
Therefore, we have a unique Nash equilibrium: 𝑝"∗ = 𝑝#∗ = 𝑐
This is an equilibrium strategy profile since neither firm can gain by
raising or lowering its price.
Corresponding equilibrium quantity is
𝑄 =𝛼−𝑝 =𝛼− 𝑐
Firms’ profits:
1
𝜋" 𝑝" = 𝑐, 𝑝# = 𝑐 = 𝑐 𝛼 − 𝑐 − 𝑐 𝛼 − 𝑐 = 0
2

Firms get zero profits in equilibrium

 Bertrand vs Cournot
• Aggregate sales are higher under Bertrand than under Cournot:
2
𝑄 = 𝛼 − 𝑐 > 𝛼 − 𝑐 = 𝑄&
%
3

• Equilibrium price is lower under Bertrand than under Cournot:

% %
1
𝑝" = 𝑝# = 𝑝 = 𝑐 < 𝛼 + 2𝑐 = 𝑝&
%
3

• Individual profits are lower under Bertrand than under Cournot:

1
𝜋! = 0 < 𝛼 − 𝑐 # = 𝜋!&
%
9
EX.5 ch.8 Price selection with differentiated products
Consider a duopoly in which two firms 𝑖 = 1,2 simultaneously and
independently select prices. The firms’ products are differentiated.
Consumers demand 𝑞" = 10 − 𝑝" + 𝑝# units produced from firm 1
and 𝑞# = 10 − 𝑝# + 𝑝" units produced from firm 2. Assume that
each firm must supply the number of units demanded. Also
assume that each firm produces at zero cost. The payoff for each
firm is equal to the firm’s profit.
a) Write the payoff functions of the firms.
b) Compute firm 𝑖’s best-response function.
c) Find the Nash equilibrium of this game.
d) Can you determine the rationalizable strategies in this game by
inspecting the graph of the best-response functions?
a) Firm 𝑖’s payoff function
𝜋! = 𝑝! 10 − 𝑝! + 𝑝'
b) Firm 𝑖’s best-response function
𝜕𝜋!
= 10 − 2𝑝! + 𝑝' = 0
𝑝!
1
𝑝! = 5 + 𝑝'
2
c) Nash equilibrium
"
𝑝! = 5 + 𝑝'
#
D " 𝑝!∗ = 𝑝'∗ = 10
𝑝' = 5 + 𝑝!
#
d) rationalizable strategies

#"
!"! #"

!"" #!

10
5

5 10 #!
 Duopoly with capacity constraints (Bertrand variation)
Two firms produce a homogeneous good and compete by selecting
prices simultaneously and independently
The firms produce at zero cost
There are 10 consumers
Each consumer would like to buy 1 unit of the good and is willing to
pay at most 𝒑 = 𝟏 for the good.
In a setting with unconstrained capacities, all of the consumers
purchase from the firm that sets the lower price, unless this price
exceeds 1 in which case no one buys. If the firms charge the same
price, then the consumers split equally between the firms. The
unique Nash equilibrium of this game without capacity constraints
has 𝒑∗𝟏 = 𝒑∗𝟐 = 𝟎 and profit is zero for both firms.
Case of capacity constraint:
each firm can produce and sell at most 𝟖 units. The firm with the
cheapest price cannot capture the entire market, while the firm with
the highest price can still capture two consumers.
With capacity constraints, the game no longer has an equilibrium in
pure strategies. No NE if 𝑝" = 𝑝# :
• If 𝑝" = 𝑝# > 0, either firm would like to undercut the other firm’s
price by an infinitesimal amount. This will always yield a higher
payoff than choosing the same price as the opponent.
• 𝑝" = 𝑝# = 0 is no longer a NE because if my opponent sets a price
of zero, my best response now is to set a price of $1. This will
ensure me a profit of $2 instead of $0, which is what I would
obtain if I had set my price to zero.
Case of capacity constraint:
What if 𝑝" ≠ 𝑝# ?
• If 𝑝" < 𝑝# ≤ 1, then it must be 𝑝# = 1, since firm 2 would have
two costumers anyway. But if 𝑝# = 1, firm 1 would like to keep
raising 𝑝" by infinitesimal amounts to become closer and closer
to $1. So, the best response by would not be well-defined.

Therefore, since there is no pure-strategy Nash equilibrium either in

case where 𝑝" = 𝑝# or where 𝑝" ≠ 𝑝# , we conclude that this game
does not possess a pure strategy Nash equilibrium.
Equilibrium
a)Same prices: if 𝑝! = 𝑝" , 𝑄 = 10 is split between the two firms so 𝑞! = 𝑞" =
5. To maximize profits, since 𝑝! , 𝑝" ∈ 0,1 , prices must be 𝑝! = 𝑝" = 1, with
𝜋! = 𝜋" = 5 (if 𝑝! = 𝑝" < 1, then profits are lower). So, 𝜋()*+, , 𝜋-)*+, = 5 is a
lower bound to profits if firms set the same price.

b)Different prices: firm 𝑖 can deviate choosing 𝑝( < 𝑝- to produce 𝑞( = 8 (if

𝑝( = 0.99 then 𝜋( = 0.99 8 = 7.92 > 5) while firm 𝑗 still sets 𝑝- = 1 with
𝜋- = 2. So, 𝜋( , 𝜋- = 2 is a lower bound to profit if firms set different prices.
.!"#$% 0
Here the price war ends when one firm sets, 𝑝( = = = 0.625 because
/! 1
now 𝜋( = 0.625 8 = 5 and 𝜋- = 0.63 2 = 1.26. Firm 𝑗 can now deviate
choosing 𝑝- = 1, with 𝜋- = 2, or, better still, 𝑝- = 0.625, with 𝜋- = 𝜋( =
0.625 5 = 3.125. The option 𝑝( = 𝑝- = 1, with 𝜋( = 𝜋- = 5 is better, so there
is no profile of prices that constitute a pure-strategy Nash equilibrium.
 Mixed-strategy equilibrium
For each firm 𝑖, 𝐹 𝑥 : 0,1 is the probability that firm 𝑖 charges a
price 𝑝! ≤ 𝑥.
If firm 𝑗 plays according to distribution F and if firm 𝑖 chooses price
𝑝! , then firm 𝑖’s expected payoff is
8𝑝! 1 − 𝐹 𝑝! + 2𝑝! 𝐹 𝑝!

Where 𝐹 𝑝! is the probability that firm 𝑗’s price is below 𝑝! , so firm

𝑖 sells to 2 consumers only, while 1 − 𝐹 𝑝! is the probability that
firm 𝑗’s price is higher than 𝑝! , and firm 𝑖 sells to 8 consumers.
Since 𝐹 1 = 1, firm 𝑖 gets a payoff of 2 by charging the price 𝑝! = 1.
Firm 𝑖’s expected payoff must be 2 over all prices in the interval 𝑝, 1 . This
leads to the following identity:
8𝑥 1 − 𝐹 𝑥 + 2𝑥𝐹 𝑥 = 2
for all 𝑥 ∈ 𝑝, 1 . Solving for F, we get:
4𝑥 − 1 4 1
𝐹 𝑥 = = −
3𝑥 3 3𝑥
This expression also reveals the value of the lower bound 𝑝 because we know
that 𝐹 𝑝 = 0 by definition. Solving this equation yields
2 ! 2 ! !
$
− $3 = 0 → $
= $3 → 𝑥=𝑝=2

In the mixed-strategy equilibrium, the firms select prices by randomizing over

!
the interval , 1 according to cumulative distribution function F. Each firm
2
gets an expected payoff of 2 in equilibrium.
A simple capacity constraint leads firms to randomize as they fight
for position in the market.
The randomization probabilities balance two forces.
By raising its price, a firm:
1) increases its profit margin

2) but decreases the probability that it will be the lower priced firm
and capture the higher quantity.