Topic 13

Strategies Over Time

(Chapter 13)
Ratna K. Shrestha
 13.1 Repeated Games
⚫ In Repeated Games actions are taken and
payoffs are received over and over again.
Oligopolistic firms (such as Intel vs. AMD) play
repeated games.
⚫ With each repetition of the Prisoners’
Dilemma, firms can develop reputations about
their behavior and study the behavior of their
competitors.
⚫ Firms choose strategies this period depending
on the rival’s behavior in the previous periods.
2
 Repeated Game
Firm 2
Low Price High Price

Low Price 10, 10 110, -50

Firm 1

High Price -50, 110 50, 50

3
 Repeated Game
⚫ A tit-for-tat strategy for repeated prisoners’
dilemma games sets cooperation in the 1st
round (High Price, High Price) then copies
the rival’s previous action in each subsequent
round.
⚫ If a rival defects from a collusive outcome at
time t, that can trigger a punishment at time
t+1. Tit for tat is one example of such a
trigger strategy.

4
 Repeated Games
⚫A Tit-for-Tat may not induce cooperation in a
repeated game if the extra profit in period t from
defection (that is, 110-50=60 in Slide #3) is greater
than the loss from the punishment in period t + 1
(=50-10 = 40).
⚫However, if the tit-for-tat strategy is modified to
extend the punishment for more than one period,
then it may ensure cooperation.
The extension should be enough to more than
compensate the one time extra profit.
 Repeated Game: Tit-for-Tat
⚫ What if the game is infinitely repeated?
Tit-for-tat strategy is rational in this case. If
a competitor undercuts its rival, it will get
high profits that month but knows the rival
will lower price next month. As a result both
will get lower profits. So it is not rational to
undercut.
 Thus such trigger strategy may lead to
cooperative Nash outcome that is best for
both. Such tacit collusions are legal.
6
 Repeated Game: Tit-for-Tat
⚫ Thus in an infinitely repeated game,
prisoners dilemma can have cooperative
outcome (but not necessarily).
⚫ Cooperation may fail due to regulation,
bounded rationality or if future profits are less
important (that is discount rate is high).
⚫ For cooperation the present value of extra
profits earned in the future from cooperation
must be higher than one time profits from
defection.
7
 Finite Game
⚫ What if the game is repeated a finite number
of times once every month?
After the last month, there is no retaliation
possible. But in the month before last
month, knowing that the rival would charge
low price in last month, it will charge low
price in month before.
Keep going backwards and see that only
rational outcome is for both firms to charge
low price every month.
8
13.2 Sequential Games
⚫ Players move in turn, responding to each
other’s actions and reactions.
Example: Stackelberg model
Responding to a competitor’s
advertisement campaign.
Entry decisions.
Responding to a regulatory policy.

9
 Sequential Game
⚫ In the product choice game, if both the firms
announce their decisions simultaneously,
there is multiple Nash equilibrium.
⚫ What if firm 1 speed up production and
introduce new cereal first?
⚫ In this sequential game, Firm 1 (the leader)
considers firm 2’s (follower) potential reaction
before making its own choice (Stackelberg
Model).
10
Payoff Matrix
Firm 2
Crispy Sweet

Crispy -5, -5 10, 20

Sweet 20, 10 -5, -5

11
 Product Choice Game in Extensive
Form
(Firm 1, Firm 2)

Crispy -5, -5
Crispy Firm 2
Sweet 10, 20
Firm 1
Crispy 20, 10 X
Sweet Firm 2
Sweet -5, -5

12
 Sequential Games
⚫ In the extensive form of this game, there is a
unique Nash equilibrium (Note: in the normal
form there is multiple Nash equilibrium).
 If Firm 1 chooses Crispy, Firm 2 chooses
Sweet (20 vs. -5).
On the other hand, if Firm 1 chooses
Sweet, Firm 2 chooses Crispy (10 vs. -5).

13
 Sequential Games
⚫ The Advantage of Moving First
 In this product-choice game, there is a
clear advantage to moving first.
 Firm 1 by choosing “sweet” can force firm 2
to choose “Crispy” so that Firm 1 can get
$20. Note that if firm 1 chooses Crispy, it
ends up with $10.
 So (Firm 1, Firm 2) = (Sweet, Crispy) is the
subgame perfect Nash equilibrium.
14
Sequential Games: United vs.
American
A, U
48
5, 5
United
48 64
4, 6
American

64 48 6, 4
United
64
1, 1

United will never choose the strategies

indicated by red lines as 5 < 6 and 1< 4.
15
 Sequential Games: United Vs.
American
⚫ We can use Backward Induction to solve for
the Nash equilibrium of a game.
⚫ United will never choose the strategies
indicated by red lines. Out of the two possible
outcomes (indicated by blue lines), American,
the leader, can make more by producing 64.
⚫ So (American, United) = (64, 48) is the Nash
equilibrium. As a result, American makes $6
and United $4.

16
 Sequential Games: United Vs.
American
⚫ If United can chose first (leader), then the
Nash equilibrium changes to (American,
United) = (48, 64) and United will make more
$64. Thus clearly the leader has an
advantage in a sequential game.
⚫ Thus when a game with multiple Nash
equilibrium in normal form (simultaneous
game) is played sequentially, we have unique
Nash equilibrium. This Nash equilibrium
changes depending on who is the leader.
17
 Sequential Games: United Vs.
American
⚫ However, if both the players have dominant
strategies in simultaneous game, then playing
the game sequentially would not change the
Nash equilibrium irrespective of who is the
leader.
⚫ Consider a Prisoner’s Dilemma game. If this
game is played sequentially, the Nash
equilibrium does not change irrespective of
who moves first.

18
 Sequential Game

Firm B
No Ad Ad

No Ad 2, 2 0, 3

Ad 3, 0 1, 1

⚫ What is the Nash equilibrium of this game if

played sequentially, instead of simultaneously?

19
 Sequential Games
A, B

NO 2, 2
Firm B
NO AD
0, 3
Firm A

3, 0
AD NO
Firm B
AD
1, 1

⚫ Will the Nash equilibrium change if Firm B

(not A) is the leader? NO.
20
 Credible Threats
⚫ In simultaneous games, why would not a firm
threaten to produce more and force the rival
to produce less?
⚫ If Firm A produces leader’s output (more) but
Firm B produces Cournot output, then Firm A
would make less profits than had it also
produced Cournot output.
⚫ Note in Cournot game, your best response is
to produce Cournot output (not more).
21
 Credible Threats
⚫ Thus such
threats of
(producing more
than Cournot
output) are not
credible in a
Cournot game.
For it to be
credible it must
be rational.
22
 Credible Threats
⚫ A player can take some action to gain
advantage in a market place (for example to
deter entry or raise price).
⚫ Such action that gives a player an advantage
must be credible, however.
⚫ It can be credible only if the player takes
such action by constraining its own behavior
(e.g., investment in a new plant). In this case
the rival can believe that the firm’s strategy is
rational.
23
 Credible Threats

⚫ Why commitment or constraining makes

threat credible?
⚫ If a general burns a bridge behind an
advancing army, then troops can only
advance and not retreat. Such army can be a
fearsome fighters.
⚫ Thus limiting the options, the army can be
formidable fighters.

24
 13.3 Entry Deterrence
⚫ An incumbent gas station (at a highway exit)
can pay $b to the landlord to have exclusive
right to be the only gas station. In this case it
can monopolize the market and earn 10 – b,
where monopoly profits = 10.
⚫ If it does not pay $b, the potential entrant can
enter in which case it has to play Cournot
game with the entrant and earn Cournot profits
Πd = $4. The entrant (with identical MC) gets 4
– F, where F is the fixed cost of entry.

25
Paying to Prevent Entry
 Entry Deterrence
⚫ Blockaded Entry: If F > 4, the potential
entrant would not enter even if the incumbent
fails to get an exclusive right. The Nash
equilibrium is (Don’t Pay, Don’t Enter).
⚫ Deterred Entry: If F < 4 and 10 – b > 4, then
Incumbent will pay $b to exclude the potential
entrant. The Nash equilibrium is (Pay, Don’t
Enter).
⚫ Accommodated Entry: If F < 4 and 10 – b <
4, the Nash equilibrium is (Don’t Pay, Enter).

27
 13.4 Investing to Deter Entry
⚫ An incumbent (monopoly) may invest in
equipment even if the cost savings may not
justify investment, if it can deter entry.
⚫ Looking at the incumbent’s choice, the
potential entrant decides whether or not to
enter the market.
⚫ In the example (next slide), if incumbent
doesn’t invest, the potential entrant enters
($4 > 0). If the incumbent invests, the entrant
doesn’t enter (0 > -1).
28
 Investing to Deter Entry
Inc, Ent
NO 10, 0
Entrant
Enter
No Inv
4, 4
Incumbent

8, 0
Invest NO
Entrant Enter
2, - 1

⚫ Out of the two possibilities (blue lines),

Incumbent is better off choosing Invest. So
(Invest, Don’t Enter) is the Nash Equilibrium.
29
 Cost & Innovation Strategies
⚫Thus a first mover firm may benefit from using a
strategy that raises its own cost but raises its rivals’
costs by more.
⚫Strategies for incumbents to raise rival’s costs:
Lobby the government for more industry
regulations that raise costs for new entrants.
Impose a switching fee to customers that take
their business elsewhere or designing products
that don’t work with the rival’s.
Use patents to block rivals’ entry.
 Quiz 1
P1, P2
Sign
15, 15
player 2

sign
Don't 40, 18
Player 1

Sign 18, 40
don't sign
player 2

Don't 30, 30

What is the Nash equilibrium?

a) (Sign, Sign) b) (Don’t, Sign) c) (sign, don’t)
d) (Don’t, Don’t) e) Multiple Nash
31
 Quiz 2
High (E, IM)
(20, 100 - Z)
In IM
Low
Entrant (E) (-10, 70)
High
out IM (0, 150 - Z)

Suppose that prior to the game, IM threatens that a low

price will be chosen if E enters. What is the minimum
value of Z that makes the threat credible?
a) 70 b) 20 c) 150 d) 30 e)100
32
 Quiz 3
(E, IM)
High (2, 3)
In Big Brew Low
(-1, 1)
Little Kona
High
out Big Brew (0, 7)

If BB can monopolize the market, it can make the

maximum profits possible. For this to happen, what is
the minimum amount BB must pay LK so that LK will
stay out of the market? a) 7 b) 3 c) 1 d) 2

33
 Clicker Q# 4

4. Given linear demand and homogeneous products,

which of the following statements is TRUE?
A. The quantity of production under Stackleberg
model is lower than that under Cournot model.
B. In the Stackelberg model, the leader makes more
profits than the follower.
C. If MC of firm 1 = $12 and that of firm 2 = $15, the
price under Bertrand model = $12.
D. The price under Betrand model is lower than that
under Cournot duopoly.
 Clicker Q# 5

A. A statement made by the

labor union that it will
select the low wage is
credible.
B. In the subgame perfect
Nash equilibrium, the
outcome is efficient.
C. Automation is a dominant
strategy for the firm.
D. If the order of moves is
reversed, the subgame
perfect Nash is efficient.
13.5 Disadvantage of Moving First

⚫ Hold Up Problem: arises when two firms

want to contract or trade with each other but
one firm must move first by making a
specific investment that can be used only in
transaction with the 2nd firm.
⚫ In this case second firm takes advantage of
the first firm.

36
Hold Up Problem

After ExxonMobil invests, it becomes a hostage of

Venezuela because the investment cannot be
transferred to other country.

37
 Minimizing Hold Ups

⚫ Contracts: The 2nd-mover firm guarantees the 1st-

mover firm that it will not be exploited.
⚫ Vertical Integration: After years of holdup problems,
for example, GM bought Fisher Body.
⚫ Quasi-Vertical Integration: GM paid for and owned
the specific capital assets of Fisher body.
⚫ Reputation Building: GM can show its record of not
acting opportunistically.
⚫ Multiple or Open Sourcing: Firms using open-source
software are not subject to a holdup problem.

38
 Moving Too Quickly

⚫ The advantage of being the 1st mover is

consumer loyalty: later entrants find it
difficult to take market share from the leader.
⚫ However, moving too quickly can be costly
as well due to the odds of miscalculating
demand and later rivals building on the
pioneer’s research to produce a superior
product.

39
 Moving Too Quickly

⚫ Moving Too Quickly Examples:

Tagamet, 1st entrant of a new class of anti-
ulcer drugs, was extremely successful
when it was introduced.
Zantac, the 2nd entrant, rapidly took the
lion’s share of the market.

40