These notes essentially correspond to chapter 14 of the text.

1 Dynamic (or sequential) games

We had been studying simultaneous games, where each …rm makes its quantity
choice or price choice without observing the other …rm’s choice. Now, we
want to extend the analysis to include sequential games, where one …rm moves
…rst, the second …rm observes this decision, and then the second …rm makes
its decision. To analyze sequential games, a structure, called a game tree,
that is slightly di¤erent than the game matrix should be used. The game tree
provides a picture of who decides when, what decisions each player makes, what
decisions each player has seen made prior to his decision, and which players see
his decision when it is made. We can start by translating the simple quantity
choice game from chapter 13 (when the …rms could each only choose to produce
a quantity of 64 or 48) into a sequential games framework.
Suppose that there are two …rms (Firm A and Firm B) engaged in compe-
tition. Firm A will choose its quantity level …rst, and then Firm B will choose
its quantity level after observing Firm A’s choice. To keep this example simple,
assume that the …rms’ quantity choices are restricted to be either 48 units or
64 units. If both …rms choose to produce 64 units, then both …rms will receive
a payo¤ of $4.1. If both …rms choose to produce 48 units, then both …rms will
receive a payo¤ of $4.6. If one …rm chooses to produce 48 units and the other
chooses to produce 64 units, the …rm that produces 48 units receives a payo¤ of
$3.8 while the …rm that produces 64 units receives a payo¤ of $5.1. This game
is sequential since Firm A chooses …rst and Firm B observes Firm A’s decision.1
While we could use the matrix (or box or normal) form of the game for the
sequential game, there is another method for sequential games that makes the
sequential nature of the decisions explicit. The method that should be used is
the game tree. A game tree consists of:

1. Nodes –places where the branches of the game tree extend from

2. Branches –correspond to the strategies a player can use at each node

3. Information sets – depict how much information the player has when he
moves (if the second player knows that he follows the …rst player but
cannot observe the …rst player’s decision then his information set is really
no di¤erent than in the simultaneous move game; however, if the second
player can observe the …rst player’s decision, then his information set has
changed)

A game tree corresponding to the quantity choice game previously described

is depicted below. The individual pieces of a game tree are also labelled. The
1 In the real-world Firm A may actually choose a quantity before Firm B, but if Firm B

gains no additional information from Firm A’s decision (such as a change in the market price),
then the game is essentially one where Firm A and Firm B choose simultaneously.

1
label for information set is pointing to the open circle that encircles the term
“Firm B”. Thus, Firm B can see how much Firm A has decided to produce. If
Firm B could not determine if Firm A decided to produce 48 or 64 units, then
Firm B would have one information set, and there would be one open circle
encircling both of Firm B’s decision nodes.

To solve sequential games we start from the end of the game and work our
way back towards the beginning. This is called backward induction. To …nd
the Nash Equilibrium (NE), we …rst determine what Firm B would do given a
quantity choice by Firm A. In this example, Firm B would choose QB = 64 as
its strategy if Firm A chose QA = 48 because $5:1 > $4:6. Also, Firm B would
choose QB = 64 if Firm A chose QA = 64 because $4:1 > $3:8. Thus, Firm B’s
strategy is: {Choose QB = 64 if Firm A chooses QA = 48; choose QB = 64 if
Firm A chooses QA = 48}. We now know what Firm B will do for any given
choice by Firm A, which means that we have an entire strategy for Firm B.
Firm A, knowing that Firm B will choose QB = 64 regardless of its quantity
choice, can now “lop o¤ the branches”that correspond to QB = 48. The reason
that Firm A can lop o¤ these branches is that it knows that it will never see
the payo¤s associated with following those branches because Firm B will never
follow them. Thus, to Firm A, the game tree looks like:

2
 I have left the payo¤s there but removed the branches. Firm A has one
decision to make, produce a quantity of 48 or a quantity of 64. If it produces
a quantity of 48, Firm B will produce 64, and Firm A will receive a payo¤ of
$3.8. If it produces a quantity of 64, Firm B will produce 64, and Firm A will
receive a payo¤ of $4.1. Since $4:1 > $3:8, Firm A will choose QA = 64. Thus,
the complete NE for this game is:
Firm A: Choose QA = 64
Firm B: Choose QB = 64 if Firm A chooses QA = 48; choose QB = 64 if
Firm A chooses QA = 48
Now, when the game is played only one payo¤ is received. To …nd this payo¤
just follow the path outlined by the NE strategy. Firm A chooses QA = 64,
and if Firm A chooses QA = 64 then Firm B chooses QB = 64, which leads to
a payo¤ of $4.1 for Firm A and $4.1 for Firm B. Notice that we didn’t use the
fact that Firm B chooses QB = 64 if Firm A chooses QA = 48 because Firm
A did not choose QA = 48. We still need to include that piece as part of our
NE strategy even though we don’t use it when we …nd the path that the game
actually follows.

2 Sequential Bertrand Game

Recall that in a Bertrand game the competing …rms choose the price that they
want to sell at in the market. The …rm with the lowest price sells the quantity
that corresponds to the entire market quantity at that price, while the …rm with
the higher price sells nothing. If the two …rms choose the same price, then each

3
…rm sells 12 the market quantity at that price. Assume that the …rm’s are iden-
tical, and that each …rm has constant M C equal to c. To make this a sequential
Bertrand game, assume that Firm A chooses its price …rst, and then Firm B
observes Firm A’s choice and sets its own price. The game tree is depicted be-
low, with a slight modi…cation. Since …rms can choose any price greater than 0
they have an in…nite amount of strategies (PA = 0; PA = 1; PA = 1:5; :::). Since
it is impossible to write down an in…nite amount of branches that correspond
to the in…nite amount of strategies we simplify the game tree by drawing two
branches corresponding to the lowest possible price (PA = 0) and the highest
possible price (PA = 1) and then connect those two branches with a dotted line
to represent the fact that there are an in…nite amount of possibilities there.2
Also note that the payo¤s have been removed as listing an in…nite amount of
payo¤s to correspond to the in…nite amount of strategies is unrealistic.

Again, to …nd the solution of this game use backward induction. We want
to …nd out what Firm B would do in response to any price choice that Firm A
could make. Suppose that Firm A sets a really high price, above the M C of c.
Firm B’s best response would be to charge a slightly lower price and capture
the entire market. Suppose that Firm A sets a really low price, less than the
M C of c. Firm B’s best response in this case is NOT to undercut Firm A. If it
undercuts Firm A then it captures the entire market, but it captures the entire
market at a price below cost which means it is making a loss, which it could
2 Technically no …rm would choose a price above a (the intercept of the inverse demand

function) as any price above this level implies that the …rm sells 0 units and thus earns 0
pro…ts.

4
avoid by not producing at all, which means that if Firm A chooses a price less
than c that Firm B should choose a price greater than Firm A. We can assume
that if Firm A chooses a price less than c that Firm B will choose to set its price
equal to c to ensure that it does not make any losses. Suppose that Firm A
chooses a price equal to the M C of c. If Firm B chooses a price below c then it
captures the entire market, but at a price less than cost, which means that it is
making a loss. Clearly, Firm B could do better if it decided to stay out of the
market. If Firm B charges a price above c then it will not earn any pro…ts as it
allows Firm A to capture the entire market. If Firm B charges a price exactly
equal to c, then it will still earn zero economic pro…t but at least it will then
produce half of the market quantity. Formalizing this thought process into a
strategy we can write down:
8
< PA " if PA > c
PB = c if PA = c
:
c if PA < c
The term " means the smallest possible amount by which Firm B can un-
dercut Firm A’s price (perhaps a penny). Firm A now knows that Firm B will
use this strategy.3 Firm A then has to decide what it will do. If it prices
below M C it will capture the entire market but will make a loss. If it prices
above M C then Firm B will undercut its price and Firm A will sell nothing. If
Firm A chooses to price at M C then it splits the market quantity with Firm B.
Thus, Firm A chooses to set PA = c, which means that Firm B will set PB = c,
which means that in the sequential Bertrand game the result is the same as in
the simultaneous Bertrand game.4

3 Sequential Quantity Game

The sequential quantity game is called a Stackelberg game, after its “creator”.
In this game one …rm chooses its quantity …rst and then the other …rm observes
this quantity decision and chooses its quantity. We will assume the linear
inverse demand function, P (Q) = a bQ, where Q = qA + qB and where …rms
costs are such that T C = c qA and M C = c. The game tree for this example
is:
3 It’s
not that Firm B tells Firm A the strategy it will use, it’s that Firm A knows the
game that will be played and can also see what Firm B’s best responses will be given Firm
A’s choice of price. Also, Firm B’s strategy could have one more tier to it. If Firm A chose
any price above the monopoly price, Firm B’s best response would be to choose the monopoly
price, not to undercut Firm A by a tiny amount. Then, for any price between the monopoly
price and MC, Firm B’s best strategy would be to undercut Firm A by the smallest possible
amount. This, however, does not e¤ect the result of the game.
4 Technically, if the price space is discrete then there is a NE where both …rms choose a price

at the lowest possible increment above M C. If c = 12, and …rms must price in increments
of pennies, then the NE result is that both …rms charge $12.01 and make very, very small
economic pro…ts. This is true of the simultaneous game as well.

5
 Notice that this is the same picture as the sequential Bertrand game, only
now the …rms are making quantity choices. Again, begin with …nding Firm B’s
strategy. When we worked the simultaneous Cournot game we found the best
response functions for each …rm. Firm B’s best response function, for a given
choice of qA , was:
a c bqA
qB =
2b
Since this problem has the same basic structure, Firm B’s best response
function is the same as it was in the Cournot game. Thus, for any choice of qA
we know the exact quantity amount that Firm B would choose. There is one
a c
slight caveat to this. If Firm A were to choose an amount of qA b , then
Firm B would choose to produce 0. The reason why is that if Firm A chooses
qA = a b c , then it is choosing to produce the competitive quantity, where the
price in the market equals marginal cost. If Firm A for some reason decides to
produce a quantity qA > a b c , then Firm A is producing a quantity such that
the price in the market is LESS than M C. In this case, Firm B would opt
out of the market and produce 0, as producing 0 ensures Firm B of receiving
0 pro…ts, while producing any positive quantity will only force the price lower
and ensure that Firm B earns a loss. To summarize, Firm B’s strategy is:
a c bqA a c
2b
if 0 qA b
qB = a c
0 if qA > b
Firm A then takes Firm B’s strategy as given. Firm A is like any other
pro…t maximizing …rm, and will set M R = M C. We know what Firm A’s M C

6
is as it is given. We need to …nd Firm A’s M R. Let’s look at the inverse
demand function, P (Q) = a bQ. We know that:

P (Q) = a bqB bqA

However, as long as Firm A does not produce more than the perfectly com-
petitive quantity then Firm B will produce qB = a c 2b2bqA . We can plug this
into the inverse demand function to …nd:

a c bqA
P (qA ) = a b bqA
2b
Simplifying:

a c bqA
P (qA ) = a bqA
2
Simplifying:
a c bqA
P (qA ) = a + + bqA
2 2 2
Combining terms:
a c bqA
P (qA ) = +
2 2 2
Now, we know that this is almost in the form of P (Q) = a bQ, which we
bqA
know has M R = a 2bQ. If we let a+c 2 = A, then P (qA ) = A 2 , which
bqA
means that M R = A 2 2 =A bqA . Thus,
a+c
MR = bqA
2
Now, setting M R = M C we have:
a+c
bqA = c
2
Or:
a+c
c = bqA
2
Or:
a c
= bqA
2
Or:
a c
= qA
2b
Thus, if Firm B uses the strategy that we found, Firm A will produce qA =
a c
2b . So the NE to the Stackelberg game is:

7
 a c
qA =
2b
a c bqA a c
2b
if 0 qA b
qB = a c
0 if qA > b
We can …nd the payo¤s to the …rms of using these strategies by plugging
qA = a2bc into Firm B’s best response function to determine how much Firm B
will produce.
a c
a c b 2b
qB =
2b
Or:
a c
a c 2
qB =
2b
Or:
a c
a c 2 + 2
qB =
2b
Or:
a c
2 2
qB =
2b
Or:
a
c
qB =
4b
Thus, if Firm A produces qA = a2bc , Firm B will produce qB = a4bc . Note
that this is NOT the NE strategy for Firm B, just what the result is of Firm B
using its NE strategy. Total market quantity is then qA + qB = a2bc + a4bc =
3 a c 3
4 b , or 4 of the perfectly competitive quantity.

3.1 Comparing the results

It’s important to compare the results of the di¤erent market models. In the
table below, I have used the values that we have been using in class, a = 120,
b = 1, and c = 12 to compare the monopoly (or cartel), Cournot, Stackelberg,
and perfectly competitive (or Bertrand, both simultaneous and sequential) out-
comes. The column for CS stands for consumer surplus and the column for T S
stands for total surplus, where total surplus is de…ned as the sum of the …rm’s
pro…ts and the consumer surplus.
Q qA qB P (Q) A B CS TS
Monopoly 54 27 27 66 1458 1458 1458 4374
Cournot 72 36 36 48 1296 1296 2592 5184
Stackelberg 81 54 27 39 1458 729 3280.5 5467.5
Bertrand 108 54 54 12 0 0 5832 5832

8
 As should be clear from the total, consumer’s are made better o¤ at the
expense of the …rms as we move down the table. It is interesting to note that
the Cournot case, with two identical …rms, is slightly less e¢ cient than the
Stackelberg case, with one large …rm and one small …rm (in terms of relative
quantities produced). This raises the question of why antitrust policy may
focus on the industry with one large …rm and one small …rm, rather than the
one with two equal-sized …rms. The reason has to do with the dynamic aspects
of the markets, which we will now discuss in the form of entry prevention by a
monopolist.

3.2 Entry prevention strategies

It is possible to model entry decisions by new …rms as well as entry prevention
strategies by incumbent …rms as sequential games. Consider the case of an
incumbent monopolist who wishes to deter a possible entrant from entering the
market. There are three possible scenarios that could occur:

1. Blockaded entry: The entrant …nds it unpro…table to enter or cannot enter

due to legal restrictions. It is possible that no strategic actions have been
made by the monopolist in this case.
2. Deterred entry: The monopolist undertakes strategic actions that make
market conditions such that the entrant will not enter the market.
3. Accommodated entry: The monopolist allows the entrant to enter the
market as it is too costly to deter the entrant.

3.2.1 Blockaded entry

We could actually model blockaded entry as a 1-player decision by the monop-
olist. Suppose that the government grants a …rm monopoly power in a market
provided that the …rm pays the government a fee. If the …rm pays the fee, F ,
then it will become a monopoly and will receive m when it is in the market.
However, if the …rm refuses to pay the fee, then a second …rm will also produce
in the market. Both …rms will then receive the Cournot pro…t, C . As a
simple game, this would look like (Note that I have also included the 2nd …rm’s
payo¤ in the game tree):

9
 We can actually solve for the amount of the fee that the government could
charge the …rm. The …rm will pay the fee if the pro…t from paying the fee and
receiving monopoly power is greater than the pro…t from not paying the fee and
receiving the 2-…rm Cournot payo¤, or:
m C
F
Using some earlier results, we know that, for a linear inverse demand function
and constant marginal costs:

2
m (a c)
=
4b
2
C (a c)
=
9b
We can plug those in to …nd:
2 2
(a c) (a c)
F
4b 9b
Or:
2 2
(a
c) (a c)
F
4b 9b
Getting a common denominator:

10
 2 2
9 (a c) 4 (a c)
F
36b 36b
Combining terms:
2
5 (a c)
F
36b
m (a c)2
Now, one little trick. We know that = 4b . The term on the
5 (a c)2
left-hand side of the equation is just 9 4b . This means that:
2
5 (a c)
F
9 4b
As long as the fee is less than 95 of the monopoly pro…t the …rm would be
willing to pay the fee. Think about that – the government can essentially re-
ceive 55% of the monopoly pro…t if it grants a …rm the right to be a monopoly
producer. This could be a decent method of generating revenue for the govern-
ment.

3.2.2 Deterred entry

There are many strategic actions an incumbent monopolist might take to deter
entry. In all cases it needs to be asked whether or not these actions are credi-
ble. If they are credible then the monopolist will achieve its goal and keep the
competitor out. If they are not credible then the monopolist has just wasted
money pursuing a strategy that the entrant will ignore. Some of those are listed
below:

1. Monopolist could make a large …xed cost (or sunk cost) investment to
appear to credibly commit to producing a large amount if a potential
competitor decides to enter.
2. Over time, the monopolist could develop a reputation as a …erce competi-
tor when someone threatens his market. This reputation could act as a
credible commitment.
3. In multi-product markets, monopolists may attempt to “…ll in all the
niches”. Consider the market for breakfast cereals. There are healthy
cereals, frosted cereals, fruit-‡avored cereals, chocolate cereals etc. If the
monopolist can …ll in all the niches before his competitors get there, then
he can capture the market.
4. The monopolist, if it has a cost advantage over the potential entrant, can
set a price below the true pro…t-maximizing price in order to keep the
potential entrant out. This will work if the monopolist can charge a price
lower than the entrant’s cost but still above his own cost.

11
 In this picture the true pro…t-maximizing price for the monopolist is above
the potential entrant’s MC. Thus, the potential entrant may decide to enter the
industry. However, if the monopolist wishes to deter entry then it could charge
a price slightly below the entrant’s MC. The monopolist will still receive positive
pro…ts since the price is above cost, but the entrant will not enter because the
price is below its cost, preserving the monopoly for the monopolist.

12