Mathematical Finance

Lecture Note 3
3. Multiperiod market models

Goal: discuss properties of models for the dy-

namics of risky and risk-free assets over several
periods.
3.1 The market model

Model assumptions:

• Time scale: discrete time scale n = 0, 1, 2, ...

• Risky assets: m risky assets with prices

 For j = 1, ..., m, Sj (0) denotes current
price of the j th risky asset, a positive
constant
  For n = 1, 2, ..., Sj (n) denotes future
price of the j th risky asset, a positive
random variable
 For each of j = 1, ..., m, Sj (n), n > 0, is
a stochastic process

• Risk-free asset: risk-free bond (or bank

account) with price A(n), a positive con-
stant for every n = 0, 1, ...

A portfolio is a vector x1(n), ..., xm(n), y(n)

 
•
indicating the units of stocks and bonds
held by an investor between times n − 1
and n.

• An investment strategy is a sequence of

portfolios x1(n), ..., xm(n), y(n) with n =
 

1, 2, ..., where
  xj (n) denotes the units invested in the
j th risky asset from time n − 1 to n
 y(n) denotes the units of the risk-free
asset held from time n − 1 to n. We
assume that xj (n) and y(n) can be ran-
dom (not known at time 0) and take
any value in R.

• The value of a strategy at time n is de-

noted by V (n) and given by
m
if n > 1

P
y(n)A(n) + xj (n)Sj (n)



j=1
V (n) = m
y(1)A(0) +



P
xj (1)Sj (0) if n = 0
j=1

We call a strategy
 
• x1(n), ..., xm(n), y(n) , n =
1, 2, ... self-nancing if
m
X
V (n) = y(n + 1)A(n) + xj (n + 1)Sj (n)
j=1
 for all n = 1, 2, ...
=⇒ value before portfolio reallocation is
the same as value after portfolio realloca-
tion.
=⇒ No consumption or injection of funds
from outside.

• A strategy is predictable if for any n =

0, 1, 2, ... the portfolio
 
x1(n + 1), ..., xm(n + 1), y(n + 1)
depends only on the asset prices up to time
n. 
=⇒ x1(n + 1), ..., xm(n + 1), y(n + 1) can


be random, but the randomness is related

to the market development only up to time
n and not later.
=⇒ investor cannot foresee the future.

A strategy x1(n), ..., xm(n), y(n) is said to

 
•
be admissible if it is self-nancing, pre-
dictable and its value is nonnegative at all
times: V (n) > 0 for all n > 0.
  in single-period models: strategy is ad-
missible if its value is nonnegative
 in multiperiod models: additional con-
ditions: self-nancing & predictable

Example 1) Which of the following strategies

are predictable? I buy tomorrow at noon 10
shares of facebook stock.
(a) if its price is below $130 at the end of
tomorrow's trading day.
Ans: No, the information is available only after
the trading decision.
(b) if its price is below $130 at the beginning
of tomorrow's trading day.
Ans: Yes, the information is available before
the trading decision.
(c) if its price is above $130 in one year.
Ans: No, the information is available only after
the trading decision.
(d) if a survey done right now shows that the
majority of the students in this class believe
that the price will be above $130 in one year.
Ans: Yes, the information is available before
the trading decision; the information relevant
for the trading decision is the outcome of the
survey and not the actual price in one year.
Example 2) Consider a single-period model
with A(0) = S(0) = 100, A(1) = 105, and
110, scenario ω1


S(1) = 100, scenario ω2


90, scenario ω3




What is the set of all admissible portfolios (x, y)?

Solution: Admissible portfolios need to satisfy
V (0) > 0 and V (1) > 0. This means

V (0) > 0 ⇐⇒ 100x + 100y > 0 ⇐⇒ y > −x,

 −18 x
 
90x + 105y > 0 y >
 
 
 21
V (1) > 0 ⇐⇒ 100x + 105y > 0 −20
⇐⇒ y > 21 x
y > −22

 


110x + 105y > 0 
21 x

Thus, the admissible portfolios are given by

−18 −22
or
 
2
(x, y) ∈ R x > 0 : y > x, x<0:y> x
21 21

Proposition 3.1 Suppose that the

 initial wealth
V (0) and a predictable sequence x1(n), ..., xm(n) ,


n = 1, 2, ... of positions in the risky assets are

known. Then there exist random variables y(n)
such that the sequence x1(n), ..., xm(n), y(n) ,


n = 1, 2, ... is a predictable self-nancing in-

vestment strategy.
m
P
V (0)− xj (1)Sj (0)
Proof: Dene y(1) = j=1
A(0)
. Then
m
P
V (1) = y(1)A(1) + xj (1)Sj (1)
j=1
 m
P
V (1)− xj (2)Sj (1)
Dene y(2) = A(1)
j=1
. Then
m
V (2) = y(2)A(2) +
P
xj (2)Sj (2) and so on.
j=1
Thus, we canm see that
P
y(n)A(n) + xj (n)Sj (n) =
j=1

m
y(n + 1)A(n) +
P
xj (n + 1)Sj (n) for
j=1
n = 1, 2, ....

So, the strategy is predictable and self-nancing.

• Remarks:
 Proposition 3.1: a self-nancing strat-
egy is determined by initial value and
position in risky assets
=⇒ investments in the risk-free asset
are made so that the strategy is self-
nancing.
  Proposition 3.1 does not say that there
exists an admissible strategy. The value
of the resulting strategy does not need
to be nonnegative.

Example 1) Consider the following two-period

model:
risk-free A(0) A(1) A(2)
assets 100 100 110
Scenarios S(0) S(1) S(2)
ω1 100 120 140
ω2 100 120 110
ω3 100 90 99
with V (0) = 100, x(1) = 1 and
2 if S(1) = 120


x(2) =
0 if S(1) = 90

(a) Determine the position in the risk-free as-

set such that the strategy is self-nancing.
(b) Is the strategy predictable?

(c) Is the strategy admissible?

Solution: (a) We start by determining y(1),

V (0) − x(1)S(0) 100 − 100
y(1) = = =0
A(0) 100
V (1) = x(1)S(1) + y(1)A(1) = S(1)

in scenarios ω1, ω2

120
=
90 in scenarioω3
V (1) − x(2)S(1)
y(2) =
A(1)

for ω1, ω2

 120−240 = −1.2
= 90−0100

100 = 0.9 for ω3
(b) Yes, by Proposition 3.1 (or check directly).
(c) Yes, it is predictable and self-nancing by
Proposition 3.1 and the value process is non-
negative:
V (0) = 100 > 0,

in scenarios ω1, ω2  > 0


120
V (1) =
90 in scenarioω3
V (2) = x(2)S(2) + y(2)A(2)

for ω1 

280 − 1.2 · 110 = 148

for ω2 > 0

= 220 − 1.2 · 110 = 88
for ω3

0 + 0.9 · 110 = 99



Example 2) If the positions in the risky assets

are deterministic, are the positions in the risk-
free asset deterministic?
Solution: No, for example, take V (0) = 100,
x(1) = 1 and x(2) = 0 in the setting of the
previous example. Similar to the calculation
before, we have
V (1) − x(2)S(1)
y(2) =
A(1)

for ω1, ω2

 120 = 1.2
= 100
 90 = 0.9
100 for ω3
which is not deterministic.
3.2 No arbitrage

Denitions:

• An arbitrage opportunity is an admissi-

ble strategy satisfying V (0) = 0, P (V (n) >
0) > 0 for some n.

• The market model is said to be arbitrage

free if there are no arbitrage opportunities.
Example 1) Consider the following market model:
risk-free A(0) A(1) A(2)
assets 100 100 110
Scenarios S(0) S(1) S(2)
ω1 100 120 125
ω2 100 120 110
ω3 100 90 99
Is there an arbitrage opportunity if

(a) there are no restrictions on short selling?

(b) short selling of the risky asset is not al-

lowed?

(c) short selling is allowed, but transaction costs

of 5% of the transaction volume apply when-
ever the stock is traded?
Solution: (a) Yes, by looking at the dierent
scenarios, we can see that there is an arbitrage
opportunity by short selling the stock at time
1 in the scenarios ω1, ω2. Indeed, at time 1 in
scenarios ω1, ω2, we have
V (1) = 0 =⇒ 100y(2) + 120x(2) = 0
y(2) = −1.2x(2),
V (2) = x(2)S(2) + y(2)A(2)

for

−7x(2) ω1
=
−22x(2) for ω
2
As arbitrage opportunity, we choose
x(1) = 0, y(1) = 0

x(2)(ω1) = x(2)(ω2) = c

y(2)(ω1) = y(2)(ω2) = −1.2c

x(2)(ω3) = y(2)(ω3) = 0
for some constant c < 0.
(b) No. To benet from the arbitrage oppor-
tunity, we need to be able to short sell the
stock.
(c) By investing or short selling the stock at
time 1, we have transactions costs
5 = 6|x(2)| for ω , ω ,
time1:|x(2)| · 120 · 100 1 2

time 2: 5 = 6.25|x(2)|
|x(2)| · 125 · 100 for ω1,
5 = 5.5|x(2)|
|x(2)| · 110 · 100 for ω2.
Total transaction costs in time 2 values are
(6 · 1.1 + 6.25)|x(2)| = 12.85|x(2)| for ω1,
(6 · 1.1 + 5.5)|x(2)| = 12.1|x(2)| for ω2.

The value at time 2 equals

V (2) = V (2) without transaction costs - trans-
action costs
−7x(2) − 12.85|x(2)| for ω

1
=
−22x(2) − 12.1|x(2)| for ω
2
We need V (2) > 0 which is no longer possible
unless x(2) = 0 because

−7x(2) − 12.85|x(2)| < 0for all x(2) 6= 0

=⇒ no arbitrage opportunity.

Example 2) We consider a single-period with

three securities whose prices are denoted by A,
S1 and S2. The risk-free bank account satises
A(0) = 1 and A(1) = 1.5. The stock prices are
given by
 3/4 for ω1


1/4 for ω2

S1(0) = 100, K1 =
−3/4 for ω3




for ω1

 3/2

for ω2

S2(0) = 150, K2 = 1/2
for ω3

−1/2



(a) Give an arbitrage opportunity.

(b) Is there still an arbitrage opportunity if
transaction costs of 4% of the transaction vol-
ume of stock 1 apply whenever stock 1 is traded?
Solution: (a) V (0) = 0 implies
100x1 + 150x2 + y = 0, thus
y = −100x1 − 150x2. This gives

V (1) = x1S1(1) + x2S2(1) + yA(1)

= x1(S(1) − 150) + x2(S(1) − 225)

for ω1

 25x1 + 150x2

for ω2

= −25x1
−125x1 − 150x2 for ω3




We need V (1) > 0 with strict inequality in at

least one scenario. We thus need
1x
25x1 + 150x2 > 0 =⇒ x2 > − 6 1

−25x1 > 0 =⇒ x1 6 0
 6 x1 .
−125x1 − 150x2 > 0 =⇒ x2 6 − 5

This means x1 6 0 and

−1
6 x 1 6 x 2 6 − 5 x1.
6

For example, we can choose x1 = −1 and x2 =

1 , which implies
2

y = −100 · (−1) − 150 · 1

2 = 25

Therefore, (x1, x2, y) = (−1, 12 , 25) is an arbi-

trage opportunity.
(b) By investing or short selling stock 1, we
have transactions costs
4
time 0: |x1| · 100 · = 4|x1|,
100
4
time 1: |x1| · 175 · = 7|x1| for ω1,
100
4
|x1| · 125 · = 5|x1| for ω2,
100
4
|x1| · 25 · = |x1| for ω3.
100
Total transaction costs in time 1 values are
(4 · 1.5 + 7)|x1| = 13|x1| for ω1,
(4 · 1.5 + 5)|x1| = 11|x1| for ω2,
(4 · 1.5 + 1)|x1| = 7|x1| for ω3.

The value at time 1 equals

V (1) = V (1) without transaction costs - trans-
action costs
for ω1

25x1 + 150x2 − 13|x1|

for ω2

= −25x1 − 11|x1|
−125x1 − 150x2 − 7|x1| for ω3




We need V (1) > 0, which implies x1 6 0 by the

second scenario so that the other two condi-
tions read as
38x1 + 150x2 > 0 and −118x1 − 150x2 > 0,

hence − 75
19 x 6 x 6 − 59 x .
1 2 75 1
For example, x1 = −1 and x2 = 1/2 satisfy
the inequalities so that there is an arbitrage
opportunity.
Example 3) Assume that
 there is a predictable,
self-nancing strategy x1(n), ..., xm(n), y(n) ,


n = 1, 2, ... so that V (0) = 0, V (i) < 0 with

positive probability and V (i + 1) > 0 with prob-
ability 1 for some i. Show that there is an
arbitrage opportunity.
Solution: Note that x1(n), ..., xm(n), y(n) , n =
 

1, 2, ... itself is not an arbitrage opportunity be-

cause V is not always nonnegative. The idea
is to make no investment until time i and then
benet from the value increase. We choose
xj (l) = 0, y(l) = 0 for l 6 i and all j,

if V (i) < 0

x (i + 1)
j
xj (i + 1) =
0 if V (i) > 0
if V (i) < 0

y (i + 1) − V (i)
j A(i)
y j (i + 1) =
0 if V (i) > 0
and put after time i + 1 everything into the
bank account. The corresponding value pro-
cess satises
V (i) − V (i) = 0 if V (i) < 0


V (i) =
0 if V (i) > 0
=⇒ strategy is self-nancing.

If V (i) > 0, we have V (i + 1) = 0 and if V (i) <

0, we obtain
m
X
V (i + 1) = xj (i + 1)Sj (i + 1)
j=1

A(i + 1)
+y(i + 1)A(i + 1) − V (i)
A(i)

A(i + 1)
=V + 1)} −
(i {z V (i)} > 0
| A(i) | {z
>0 <0
=⇒ strategy is an arbitrage opportunity.
Example 4) Consider a binomial tree model
with d < r < u, but rather than being indepen-
dent, the one-step returns K(n) satisfy
if K(n − 1) = d for all n > 2

u
K(n) =
d if K(n − 1) = u
Does there exist an arbitrage opportunity?
Solution: Yes, we can benet from this be-
cause we know the returns in advance. For
some n > 2, choose
 1 if K(n − 1) = d


x(n) =
−1 if K(n − 1) = u

if K(n − 1) = d

− S(n−1)

A(n−1)
y(n) =
 S(n−1)

A(n−1)
if K(n − 1) = u
so that V (n − 1) = 0 and
A(n)
V (n) = S(n) − S(n − 1)
A(n − 1)
 = S(n − 1)(u − r) > 0 if K(n − 1) = d,
A(n)
V (n) = −S(n) + S(n − 1)
A(n − 1)

= S(n − 1)(r − d) > 0if K(n − 1) = u,

=⇒ there is an arbitrage opportunity.

Application to the binomial tree model:

Theorem 3.2: For the binomial tree model,
the following are equivalent:
(1) the model is arbitrage free,
(2) d < r < u,
(3) there is a risk-neutral probability.
Proof: In a single-period model, this follows
from Proposition 1.1 by using that we can mul-
tiply prices of the risky asset by A(0)
S(0)
without
changing the model: set
A(0) A(0)
S(0) = S(0) = A(0), S(1) = S(1),
S(0) S(0)
 A(0)
x= x, then xS(0) = xS(0), xS(1) = xS(1).
S(0)
Proposition 1.1 yields
A(0) d A(0) u
arbitrage free ⇐⇒ S < A(1) < S
S(0) S(0)

1 A(1) 1
⇐⇒ Sd < < Su
S(0) A(0) S(0)

⇐⇒ d < r < u
In a multiperiod model, consider an admis-
sible strategy with V (0) = 0. Then there
exists an arbitrage opportunity ⇐⇒ there is
n : P (V (n) > 0) > 0
⇐⇒ there is n : P (V (n) > 0) > 0 and
V (n − 1) = 0
⇐⇒ there is a one-step subtree with
V (n − 1) = 0 at its root and V (n) > 0 for at
least one node
⇐⇒ a one-step subtree has arbitrage
⇐⇒ d > r or u 6 r.
"(2) ⇐⇒ (3)": p∗ = u−d
r−d

hence r = p∗u + (1 − p∗)d and thus

d < r < u ⇐⇒ d < p∗u + (1 − p∗)d < u

⇐⇒ 0 < p∗(u−d) < u−d ⇐⇒ 0 < p∗(u−d) < u−d

⇐⇒ 0 < p∗ < 1.
The fundamental theorem of asset pricing

Theorem 3.3 The market model admits no

arbitrage if and only if there exists a probability
P∗ such that P∗(ω) > 0 for every scenario ω
and the discounted stock prices S̃j (n) = SA(n)
j (n)

satisfy
E∗[S̃j (n + 1) S(n)] = S̃j (n) (3.1)
for all j = 1, ..., m and all n = 0, 1, 2, ...
Proof: Those of you who are interested can
click here (but you are not expected to know
the proof for this class).
Note:

• Since A(n + 1) is deterministic, we can

write (3.1) as
A(n + 1)
E∗[Sj (n + 1) S(n)] = Sj (n)
A(n)
 • The fundamental theorem of asset pricing
says that no arbitrage exists if and only
if discounted prices are martingales under
some equivalent probability.

Example 1) Consider the following two-period

model:
risk-free A(0) A(1) A(2)
assets 100 100 110
Scenarios S(0) S(1) S(2)
ω1 100 120 140
ω2 100 120 110
ω3 100 90 99
with V (0) = 100, x(1) = 1 and
if S(1) = 120

2
x(2) =
0 if S(1) = 90

Is the model arbitrage free? If so, determine

risk-neutral probabilities.
Solution: The tree to this model is

The equations for a risk-neutral probability are

120 90 100
p∗ + (1 − p∗) =
100 100 100
140 110 120
q∗ + (1 − q∗) =
110 110 100
99 90
=
110 100
which are equivalent to
0.3p∗ = 0.1, 30q∗ = 22,
hence p∗ = 1/3 and q∗ = 11/15. A risk-neutral
probability P∗ is given by
11
P∗(ω1) = p∗q∗ =
45
 4
P∗(ω2) = p∗(1 − q∗) =
45
2
P∗(ω3) = 1 − p∗ =
3
=⇒ there exists a unique risk-neutral probabil-
ity P∗ = ( 11 , 4 , 2 ), and the model is arbitrage
45 45 3
free.
Example 2) Consider the similar two-period
model:
risk-free A(0) A(1) A(2)
assets 100 100 110
Scenarios S(0) S(1) S(2)
ω1 100 120 140
ω2 100 120 110
ω3 100 90 99
ω4 100 80 88
Is this model arbitrage free? If so, determine
the risk-neutral probabilities.
Solution: The tree to this model is

The equations for a risk-neutral probability are

120 90 80 100
p∗ + r∗ + (1 − p∗ − r∗) =
100 100 100 100
140 110 120
q∗ + (1 − q∗) =
110 110 100
99 90
=
110 100
88 80
=
110 100
which are equivalent to
11
q∗ = , r∗ = 2 − 4p∗
15
A risk-neutral probability P∗ is given by
11
P∗(ω1) = p∗q∗ = p∗ > 0 =⇒ p∗ > 0
15
4
P∗(ω2) = p∗(1 − q∗) = p∗ > 0 =⇒ p∗ > 0
15
1
P∗(ω3) = r∗ = 2 − 4p∗ > 0 =⇒ p∗ <
2
1
P∗(ω4) = 1 − p∗ − r∗ = 3p∗ − 1 > 0 =⇒ p∗ >
3
=⇒ the risk- risk-neutral probabilities are
11 4
P∗ = ( p∗ , p∗, 2 − 4p∗, 3p∗ − 1),
15 15
where 1 < p < 1
3 ∗ 2 and the model is arbitrage
free.
3.4 Derivative securities

Denition:

• Derivative securities are nancial rights

or obligations that depend on the prices of
other securities, called underlying securi-
ties.
Examples: call and put options, with the
stock being the underlying security

• Derivative securities are also called contin-

gent claims because the values are contin-
gent on the underlying securities

Extending the market model:

• Similarly as in Section 1.4: consider an ex-

tended model for the derivative pricing
• In addition to the m primary securities and
the risk-free asset, we consider k derivative
securities whose prices at time n are D1(n),
. . . , Dk (n)

• An investment strategy is a sequence of

vectors
 
x1(n), ..., xm(n), y(n), z1(n), ..., zk (n) ,
n=1,2,... where
 xj (n) denotes the units invested in the
j th primary security from time n − 1 to
time n
 y(n) denotes the units of the risk-free
asset held from time n − 1 to time n
 zi(n) denotes the units invested in the
ith derivative security from time n − 1 to
time n.
 • The value of the strategy is given by
m
X k
X
V (n) = xj (n)Sj (n) + y(n)A(n) + zi (n)Di (n)
j=1 i=1

for n > 1. The initial value is

m
X k
X
V (0) = xj (1)Sj (0) + y(1)A(0) + zi (1)Di (0)
j=1 i=1

• Admissible strategies and arbitrage oppor-

tunities are dened similar to Section 3.1.

For the pricing of derivative securities, the fol-

lowing version of the fundamental theorem of
asset pricing is helpful.
Theorem 3.4 The extended market model ad-
mits no arbitrage if and only if there exists a
probability P∗ such that P∗(ω) > 0 for every
scenario ω and the discounted prices S̃j (n) =
Sj (n)
A(n)
and D̃i (n) = Di (n)
A(n)
satisfy
E∗[S̃j (n + 1) S(n)] = S̃j (n) (3.2)

E∗[D̃i(n + 1) S(n)] = D̃i(n) (3.3)

for all j = 1, ..., m, i = 1, ..., k and all n =

0, 1, 2, ...

Note:

• Since A(n + 1) is deterministic, we can

write (3.3) as
A(n + 1)
E∗[Di(n + 1) S(n)] = Di(n)
A(n)

• Theorem 3.4 is useful in pricing derivatives.

  If there is a unique P∗ satisfying (3.2) or
if there are several P∗ with all having the
same E∗(Di(n)), the unique fair price of
the ith derivative security is
A(0)
Di(0) = E∗(Di(n))
A(n)

 If there are several P∗ satisfying (3.2)

with dierent E∗(Di(n)), there is an in-
terval of fair prices for the ith derivative
security: Di(0) takes a value in
 A(0) A(0) 
min E∗(Di(n)), max E∗(Di(n))
P∗ A(n) P∗ A(n)

Example 1) Consider the two-period model

discussed earlier:
 risk-free A(0) A(1) A(2)
assets 100 100 110
Scenarios S(0) S(1) S(2)
ω1 100 120 140
ω2 100 120 110
ω3 100 90 99
with V (0) = 100, x(1) = 1 and
if S(1) = 120

2
x(2) =
0 if S(1) = 90

Consider a European put option with strike

price 110 in this model. Calculate the arbitrage-
free prices at times 0 and 1.
Solution: Using the risk-neutral probabilities
calculated earlier, the arbitrage-free price at
time 1 is
 A(1)
D(1) = E∗(D(2)|S(1))
A(2)
for ω1, ω2

0
=
 100 · 11 = 10 for ω
110 3
We can calculate the arbitrage-free price at
time 0 either by
A(0)
D(0) = E∗(D(1)|S(0)) = E∗(D(1))
A(1)
1 2
= · 0 + · 10 ≈ 6.667
3 3

Or, alternatively,

A(0) A(0)
D(0) = E∗(D(2)|S(0)) = E∗(D(2))
A(2) A(2)
100 2
= · · 11 ≈ 6.667
110 3
Example 2) Consider the similar two-period
model:
risk-free A(0) A(1) A(2)
assets 100 100 110
Scenarios S(0) S(1) S(2)
ω1 100 120 140
ω2 100 120 110
ω3 100 90 99
ω4 100 80 88
Consider a European call option with strike
price 110 in this model. Calculate the arbitrage-
free prices at times 0 and 1.
Solution: Only in the scenario ω1, the option
has a positive payo. The price at time 1 is
A(1)
D(1)(ω1,2) = E∗(D(2)|S(1))(ω1,2)
A(2)
100 11 4
= ·( · 30 + · 0) = 20
110 15 15
D(1)(ω3,4) = 0 At time 0, we have
A(0)
D(0) = E∗(D(1)) = 20p∗
A(1)
for 1 < p < 1.
3 ∗ 2 So,
D(0) ∈ (6.667, 10)

Example 3) Consider the same example as

before, but take as payo
D(2) = max{S(0), S(1), S(2)}.

Solution: From
for ω1


 140
for ω2


120

D(2) =

 100 for ω3
for ω4


100


we conclude that
100 ( 11·140 + 4·120 ) = 122.4242 for ω1, ω2


 110 15 15
for ω3

100
D(1) = 110 · 100 = 90.9091
for ω4

 100
110 · 100 = 90.9091

The prices at time 0 are
A(0)
D(0) = E∗(D(1))
A(1)
= 122.4242p∗ + 90.9091(2 − 4p∗)
+ 90.9091(3p∗ − 1)
= 90.9091 + 31.5151p∗
for 1 < p < 1.
3 ∗ 2 So,
D(0) ∈ (101.4141, 106.6667)

Example 4) Consider a single-period model

consisting of three securities: the bank ac-
count whose price process is A(0) = A(1) = 1,
and two stocks with prices
3 for ω1


S1(1) = 1 for ω2

S1(0) = 2,
5 for ω3




for ω1

9

for ω2

S2(0) = 7, S2(1) = 5
for ω3



13
What is the fair price at time 0 of the payo
max{3S1(1), S2(1)} at time 1?

Solution: We can calculate the risk-neutral

probabilities for this model to be
3 1 1 1
P∗ = (p∗, − p∗, − p∗)
4 2 4 2
for 0 < p∗ < 21 .
Hence, the time-0 price of max{3S1(1), S2(1)}
is
A(0)
D(0) = E∗(max{3S1(1), S2(1)})
A(1)
= E∗(max{3S1(1), S2(1)})
+ 90.9091(3p∗ − 1)
3 1 1 1
= 9p∗ + 5( − p∗) + 15( − p∗)
4 2 4 2
= −p∗ + 7.5
for 0 < p∗ < 12 .
Thus, the time-0 fair prices are D(0) ∈ (7, 7.5).
Bibliography

Marek Capinski and Tomasz Zastawniak, Math-

ematics for Finance, An Introduction to Finan-
cial Engineering
John C. Hull, Options, Futures, and Other
derivatives, 9th Edition, Pearson