Chapter 4

Multi-asset options
In this chapter we introduce the idea of higher dimensionality by describing
the Black-Scholes theory for options on more than one underlying asset.
This theory is perfectly straightforward; the only new idea is that correlated
random walks and the corresponding multifactor version of Ito Lemma.
4.1 Pricing model
4.1.1 Two-asset options
Consider a European option whose payo, denoted by f(S
1
, S
2
), depends on
two assets S
1
and S
2
. The basic building block for option pricing with one
underlying is the lognormal random walk
dS
S
= dt + dW.
This is readily extended to a world containing two assets via models for each
underlying
dS
1
S
1
=
1
dt +
1
dW
1
dS
2
S
2
=
2
dt +
2
dW
2
As before, we can think of dW
i
, i = 1, 2 as a random number drawn from a
Normal distribution with mean zero and standard deviation dt
1/2
so that
E(dW
i
) = 0 and E[dW
2
i
] = dt
33
34 CHAPTER 4. MULTI-ASSET OPTIONS
but the random numbers dW
1
and dW
2
are correlated:
E[dW
1
dW
2
] = dt
Here is the correlation coecient between the two random walks..
Let V (S
1
, S
2
, t) be the option value. Since there are two sources of
uncertainty, we construct a portfolio of one long option position, two short
positions in some quantities of underlying assets:
= V
1
S
1

2
S
2
.
Consider the increment
d = dV
1
dS
1

2
dS
2
Here we need the Ito Lemma involving two variables.
dV =
_
V
t
+
1
2

2
1
S
2
1

2
V
S
2
1
+
1

2
S
1
S
2

2
V
S
1
S
2
+
1
2

2
2
S
2
2

2
V
S
2
2
_
dt
+
V
S
1
dS
1
+
V
S
2
dS
2
Actually the two dimensional Ito Lemma can be derived by using Taylor
series and the rules of thumb: dW
2
i
= dt, i = 1, 2, and dW
1
dW
2
= dt.
Taking
1
=
V
S
1
and
2
=
V
S
2
to eliminate risk, we then have
d =
_
V
t
+
1
2

2
1
S
2
1

2
V
S
2
1
+
1

2
S
1
S
2

2
V
S
1
S
2
+
1
2

2
2
S
2
2

2
V
S
2
2
_
dt.
Then the portfolio is riskless and then earn riskless return, namely
d = r = r(V
V
S
1
S
1

V
S
2
S
2
)dt.
So we arrive at an equation
V
t
+
1
2

2
1
S
2
1

2
V
S
2
1
+
1

2
S
1
S
2

2
V
S
1
S
2
+
1
2

2
2
S
2
2

2
V
S
2
2
+rS
1
V
S
1
+rS
2
V
S
2
rV = 0.
(4.1)
The solution domain is {S
1
> 0, S
2
> 0, t [0, T)}, and the nal condition
is
V (S
1
, S
2
, T) = f(S
1
, S
2
) (4.2)
(4.1-4.2) form a complete model. Well-known payos are the following:
4.1. PRICING MODEL 35
f(S
1
, S
2
) =

(max(S
1
, S
2
) X)
+
, maximum call
(X max(S
1
, S
2
))
+
, maximum put
(min(S
1
, S
2
) X)
+
, minimum call
(X min(S
1
, S
2
))
+
, minimum put
(S
1
S
2
X)
+
, spread option
(4.3)
If the assets pay continuous dividends, then (4.1) is replaced by
V
t
+
1
2

2
1
S
2
1

2
V
S
2
1
+
1

2
S
1
S
2

2
V
S
1
S
2
+
1
2

2
2
S
2
2

2
V
S
2
2
+(rq
1
)S
1
V
S
1
+(rq
2
)S
2
V
S
2
rV = 0.
(4.4)
where q
1
and q
2
are dividend yields of two assets, respectively.
4.1.2 American feature:
Suppose that the option can be exercised early receiving the payo. Then
the pricing model is
min {LV, V f(S
1
, S
2
)} = 0
V (S, T) = f(S
1
, S
2
)
where L =

t
+
1
2

2
1
S
2
1

2
S
2
1
+
1

2
S
1
S
2

2
S
1
S
2
+
1
2

2
2
S
2
2

2
S
2
2
+(r q
1
)S
1

S
1
+
(r q
2
)S
2

S
2
r
4.1.3 Exchange option: similarity reduction
An exchange option gives the holder the right to exchange one asset for
another. The payo for this contract at expiry is (S
1
S
2
)
+
. So the nal
condition is
V (S
1
, S
2
, T) = (S
1
S
2
)
+
.
The governing equation is still (4.1).
This contract is special in that there is a similarity reduction. Let us
postulate that the solution takes the form
V (S
1
, S
2
, t) = S
2
H(, t),
where the new variable is
=
S
1
S
2
.
36 CHAPTER 4. MULTI-ASSET OPTIONS
If this is the case, then instead of nding a function V of three variables, we
only need nd a function H of two variables, a much easier task.
It follows
V
S
1
= S
2
H

1
S
2
=
H

,

2
V
S
2
1
=

2
H

2
1
S
2
,
V
S
2
= H + S
2
H

S
1
S
2
2
_
= H
H

2
V
S
2
2
=
H

S
1
S
2
2
_

S
1
S
2
2
H

2
H

2
_

S
1
S
2
2
_
=
1
S
2

2
H

2
,

2
V
S
1
S
2
=
S
1
S
2
2

2
H

2
,
V
t
= S
2
H
t
.
The partial dierential equation now becomes
H
t
+
_
1
2

2
1

2
+
1
2

2
2

2
_

2
H

2
+(r q
1
)
H

+(r q
2
)
_
H
H

_
rH = 0
or
H
t
+
1
2

2
H

2
+ (q
2
q
1
)
H

q
2
H = 0,
where

=
_

2
1
2
1

2
+
2
2
. This equation is just the Black-Scholes
equation for a single stock with q
2
in place of r, q
1
in place of the divi-
dend yield on the single stock and with a volatility of

. Note that the nal

1
) S
2
e
q
2
(Tt)
N(d

2
).
where
d

1
=
log(S
1
/S
2
) + (q
2
q
1
+
1
2

2
)(T t)

T t
, and d

2
= d

T t
Remark 9 An exchange option is a kind of spread option with X = 0. If
X = 0, the similarity reduction doesnt work because the payo cannot be
reduced to a function of and t.
4.1. PRICING MODEL 37
4.1.4 Options on many underlyings
Options with many underlyings are called basket options, options on baskets
or rainbow options. We now extend the two-asset option pricing model to
a general case. Suppose
dS
i
=
i
S
i
dt +
i
S
i
dW
i
.
Here S
i
is the price of the ith asset, i = 1, 2, ..., n, and
i
and
i
are the
drift and volatility of that asset respectively and dW
i
is the increment of a
Brownian motion. We can still continue to think of dW
i
as a random number
drawn from a Normal distribution with mean zero and standard deviation
dt
1/2
so that
E(dW
i
) = 0 and E(dX
2
i
) = dt
and the random numbers dW
i
and dW
j
are correlated:
E[dW
i
dW
j
] =
ij
dt,
here
ij
is the correlation coecient between the ith and jth random walks.
The symmetric matrix with
ij
as the entry in the ith row and jth column
is called the correlation matrix. For example, if we have four underlyings
n = 4 and the correlation matrix will look like this:
D =

1
12

13

14

21
1
23

24

31

32
1
34

41

42

43
1

Note that
ii
= 1 and
ij
=
ji
. The correlation matrix is positive denite,
so that y
T
Dy 0.
To be able to manipulate functions of many random variables we need
a multidimensional version of Itos lemma. If we have a function of the
variables S
1
, S
2
, ..., S
n
and t, V (S
1
, S
2
, ..., S
n
, t), then
dV =

V
t
+
1
2
n

i=1
n

j=1

ij
S
i
S
j

2
V
S
i
S
j

dt +
n

i=1
V
S
i
dS
i
.
We can get to this same result by using Taylor series and the rules of thumb:
dW
2
i
= dt and dW
i
dW
j
=
ij
dt.
The pricing model for basket options is straightforward. Still set up a
portfolio consisting of one basket option, and short a number
i
of each of
38 CHAPTER 4. MULTI-ASSET OPTIONS
the asset S
i
, employ the multidimensional Itos Lemma, take
i
=
V
S
i
to
eliminate the risk, and set the return of the portfolio equal to the risk-free
rate. We are able to arrive at
V
t
+
1
2
n

i=1
n

j=1

ij
S
i
S
j

2
V
S
i
S
j
+
n

i=1
(r q
i
)S
i
V
S
i
rV = 0.
Here q
i
is the dividend yield on the ith asset. The nal condition is
V (S
1
, S
2
, ..., S
n
, t) = f(S
1
, S
2
, ..., S
n
)
The analytic solution to the above model is available, but involves multiple
integral, as in the case of two-asset options. (See Page 154, Wilmott (1998))
4.2 Quantos
There is one special, and very important type of multi-asset option. This is
the cross-currency contract called a quanto. The quanto has a payo dened
with respect to an asset or an index (or an interest rate) in one country, but
then the payo is converted to another currency payment. The general form
of its payo can be expressed as
f (S
$
, S) .
Here S
$
is the exchange rate between the domestic currency and the
foreign currency (for example, dollar-yen rate, number of dollars per yen),
and S is the level of some foreign asset (for example, the Nikkei Dow index).
Note that the quanto contract is measured in domestic currency, but S is in
foreign currency. So this contract is exposed to the exchange rate and the
asset.We assume
dS
$
=
$
S
$
dt +
$
S
$
dW
$
and dS = Sdt + SdW
with a correlation coecient between them.
Let V (S
$
, S, t) be the quanto option value in US dollar. Construct a
portfolio consisting of the quanto, hedged with the foreign currency and the
asset:
= V (S
$
, S, t)
$
S
$
SS
$
.
Note that every term in this equation is measured in domestic currency
(dollar).
$
is the number of the foreign currency (yen) we hold short, so

$
S
$
is the dollar value of that yen. Similarly, with the term SS
$
we
4.2. QUANTOS 39
have converted the yen-denominated index S into dollars, is the amount
of the index held short.
The change in the value of the portfolio is due to the change in the value
of its components and the interest received on the yen:
dV =
_
V
t
+
1
2

2
$
S
2
$

2
V
S
2
$
+
$
S
$
S

2
V
S
$
S
+
1
2

2
S
2

2
V
S
2
_
dt
+
V
S
$
dS
$
+
V
S
dS

$
dS
$

$
S
$
r
f
dt
S
$
dS SdS
$

$
SS
$
dt
=
_
V
t
+
1
2

2
$
S
2
$

2
V
S
2
$
+
$
S
$
S

2
V
S
$
S
+
1
2

2
S
2

2
V
S
2

$
SS
$

$
S
$
r
f
_
dt
+
_
V
S
$

$
S
_
dS
$
+
_
V
S
S
$
_
dS,
where the term
$
S
$
r
f
dt is the interest received by the yen holding, and

$
SS
$
dt is due to the increment of the product SS
$
. We now
choose
=
1
S
$
V
S
and
$
=
V
S
$

S
S
$
V
S
to eliminate the risk in the portfolio. Setting the return on this riskless
portfolio equal to the US risk-free rate of interest r, since is measured
entirely in dollars, yields
V
t
+
1
2

2
$
S
2
$

2
V
S
2
$
+
$
S
$
S

2
V
S
$
S
+
1
2

2
S
2

2
V
S
2
dt +(rr
f
)S
$
V
S
$
+(r
f

$
)S
V
S
rV = 0.
(4.5)
This completes the formulation of the pricing equation. The equation is
valid for any contract with underlying measured in one currency but paid
in another. The nal conditions on t = T :
V (S
$
, S, T) = f(S
$
, S).
Notice that these parameters correspond to two-asset options with con-
tinuous dividend payments (i.e. Eqn (4.4)), where under the risk-neutral
world, the underlying assets follow
dS
1
S
1
= (r q
1
)dt +
1
dW
1
40 CHAPTER 4. MULTI-ASSET OPTIONS
dS
2
S
2
= (r q
2
)dt +
2
dW
2
with dt = E(dW
1
dW
2
). For quanto options (i.e. Eqn (4.5)), the underlyings
follow in the risk-neutral world
dS
$
S
$
= (r r
f
)dt +
$
dW
$
dS
S
= (r
f

$
)dt + dW
= (r (r r
f
+
$
))dt + dW.
with dt = E(dW
$
W). Therefore, in this case, q
1
= r
f
and q
2
= rf
f
+
$
.
4.3 Numerical Methods
4.3.1 *Binomial tree methods
Suppose (S
1
, S
2
) will move to (S
1
u
1
, S
2
u
2
) with probability p
1
, (S
1
u
1
, S
2
d
2
)
with probability p
2
, (S
1
d
1
, S
2
u
2
) with probability p
3
,and (S
1
d
1
, S
2
d
2
) with
probability p
4
after the next timestep. Then the binomial model for two-
asset options is
V (S
1
, S
2
, t) = e
rt
[p
1
V (S
1
u
1
, S
2
u
2
, t + t)
+p
2
V (S
1
u
1
, S
2
d
2
, t + t)
+p
3
V (S
1
d
1
, S
2
d
2
, t + t)
+p
4
V (S
1
d
1
, S
2
u
2
, t + t)]
where p
i
for i = 1, 2, 3, 4, u
i
, d
i
for i = 1, 2 are chosen to be consistent with
the continuous-time model. One choice for these parameters is given as
follows
u
i
= e

t
, d
i
=
1
u
i
for i = 1, 2.
p
1
=
1
4

1 +

r q
1


2
1
2

1
+
r q
2


2
2
2

t +

p
2
=
1
4

1 +

r q
1


2
1
2

r q
2


2
2
2

p
3
=
1
4

1 +

r q
1


2
1
2

r q
2


2
2
2

t +

4.3. NUMERICAL METHODS 41

1 +

r q
1


2
1
2

1
+
r q
2


2
2
2

.
We refer interested students to Kwok (1998) [pp 207-208] for derivation of
the above parameters. It should be pointed out that we can also use the
nite dierence method to determine these parameters.
4.3.2 Monte-Carlo simulation
The amount of computation of BTM grows exponentially with the number
of underlyings. We will have to give up BTM if the number of underly-
ings is greater than 3, and instead employ Monte-Carlo simulation which is
relatively more ecient as the number of underlyings increases.
Monte-Carlo simulation is based on the risk-neutral valuation result.
The expected payo in a risk-neutral world is calculated using a sampling
procedure. It is then discounted at the risk-free interest rate.
Suppose in a risk-neutral world
dS
i
=
i
S
i
dt +
i
S
i
dW
i
, (1 i n)
As in the single-variable case, the life of the derivative must be divided into
N subintervals of length t. The discrete version of the process for S
i
is
then
S
i
(t + t) S
i
(t) =
i
S
i
t +
i
S
i

t, (4.6)
where
i
is a random sample from a standard normal distribution. The
coecient of correlation between
i
and
j
is
ij
for 1 i, j n. One
simulation trial involves obtaining N samples of the
i
(1 i n) from a
multivariate standardized normal distribution. These are substituted into
equation (4.6) to produce simulated paths for each S
i
and enable a sample
value for the derivative to be calculated.
Note that correlated samples
i
(1 i n) from standard normal dis-
tributions are required. We only give a procedure for n = 2. For n 3,
we refer interested readers to Appendix. Independent samples x
1
and x
2
from a univariate standardized normal distribution are easily obtained. The
required samples
1
and
2
are then calculated as follows:

1
= x
1

2
= x
1
+
_
1
2
x
2
where is the coecient of correlation.
42 CHAPTER 4. MULTI-ASSET OPTIONS
Remark 10 (1) At each time step, we need to nd
i
(1 i 2).
(2) The number of simulation trials M carried out depends on the accu-
racy required. In general, we take M = 5000 or 10000.
Remark 11 The amount of computation of Monto-Carlo simulation grows
only linearly with the number of underlyings. The main drawback of Monte
Carlo simulation is that it cannot easily handle situations where there are
early exercise opportunities.
4.3.3 *Generation of correlated samples
Consider the situation where we require n correlated samples from normal
distributions where the coecient of correlation between sample i and sam-
ple j is
ij
. We rst sample n independent variables x
i
(1 i n), from
univariate standardized normal distributions. The required samples are
i
(1 i n), where

i
=
i

k=1

ik
x
k
.
For
i
to have the correct variance and the correct correlation with the
j
(1 j n), we must have
i

k=1

2
ik
= 1
and, for all j i,
j

k=1

ik

jk
=
ij
.
The rst sample,
1
, is set equal to x
1
. These equations for the

s can be
solved so that
2
is calculated from x
1
and x
2
;
3
is calculated from x
1
, x
2
and x
3
; and so on. The procedure is known as the Cholesky decomposition.
For example, when n = 3,

1
= x
1

2
=
21
x
1
+
_
1
2
21
x
2

3
=
31
x
1
+

23

31

21
_
1
2
21
x
2
+

1 + 2
23

21

31

2
21

2
31

2
23
1
2
21
x
3