Stochastic Calculus for Finance II

-
some Solutions to Chapter V
Matthias Thul

January 4, 2012
Exercise 5.1
(i) Let f(t, x) = S(0)e
x
. We have
f
t
= 0,
f
x
= f(t, x)

2
f
x
2
= f(t, x)
and
dX(t) =
_
(t) R(t)
1
2

2
(t)
_
dt + (t)dW(t)
(dX(t))
2
=
2
(t)dt
By the It o formula, the dierential of the discounted stock price D(t)S(t) is given
by
d (D(t)S(t)) = df(t, X(t))
= D(t)S(t)dX(t) +
1
2
D(t)S(t) (dX(t))
2
= ((t) R(t)) d(t)S(t)dt + (t)D(t)S(t)dW(t) (q.e.d.)
(ii) We rst note that the cross variation dD(t)dS(t) = 0, since D(t) is a nonrandom
function of time and thus has zero quadratic variation. We thus obtain

The author can be contacted via <<firstname>>.<<lastname>>@googlemail.com and

http://www.matthiasthul.com.
1
d (D(t)S(t)) = R(t)D(t)S(t)dt + (t)D(t)S(t)dt + (t)D(t)S(t)dW(t)
= ((t) R(t)) d(t)S(t)dt + (t)D(t)S(t)dW(t) (q.e.d.)
Exercise 5.2 (State Price Density Process)
This assertion follows from Equation (5.2.30) and Lemma 5.2.2.
D(t)V (t) =

E[ D(T)V (T)| F(t)] =
1
Z(t)
E[ D(T)Z(T)V (T)| F(t)]
and thus
D(t)Z(t)V (t) = E[ D(T)Z(T)V (T)| F(t)] . (q.e.d.)
Exercise 5.3
(i) Dierentiating inside the expected value and applying the chain rule yields
c
x
(0, x) =

E
_
e
rT
I
{
xexp
{

W(T)+
(
r
1
2

2
)
T
}
>K
}
exp
_

W(T) +
_
r
1
2

2
_
T
__
=

E
_
I
{S(T)>K}
exp
_

W(T)
1
2

2
T
__
.
Here, we have dened S(t) by
S(t) = x exp
_

W(t) +
_
r
1
2

2
_
t
_
(ii) Let

P be a probability measure equivalent to

P and let Z(t) be a Radon-Nikod ym.
By Lemma 5.2.1, By Lemma 5.2.1, the expected value in of an F(t)-measureable
random variable Y satises

E[Y (t)] =

E[Z(t)Y (t)]
We note that I
{S(T)>K}
is F(T)-measurable and can thus write
2

P(S(T) > K) =

E
_
I
{S(T)>K}

=

E
_
Z(T)I
{S(T)>K}

=

E
_
I
{S(T)>K}
exp
_

W(T)
1
2

2
T
__
= c
x
(0, x) (q.e.d.) (1)
Here, we have dened Z(t) to be
Z(t) = exp
_

W(t)
1
2

2
t
_
.
By Girsanovs theorem (Theorem 5.2.3), this is just the Radon-Nikod ym derivative
process that renders a

P-Brownian motion into a

P-Brownian motion if

W(t) =

W(t) t.
Thus, the Z(t) fullls all required properties such that the change of measure in
Equation (1) is dened.
(iii) Substituing for S(T) and using

W(T) =

W(T) + T yields

P(S(T) > K) =

P
_
x exp
_

_

W(T) + T
_
+
_
r
1
2

2
_
T
_
> K
_
=

P
_

W(T) +
_
r +
1
2

2
_
T > ln
_
K
x
__
=

P
_

W(T)

T
>
ln
_
K
x
_

_
r +
1
2

2
_
T

T
_
=

P
_

W(T)

T
< d
+
(T, x)
_
Since

W(T)

T
N(0, 1), we nally obtain

P(S(T) > K) = N (d
+
(T, x)) (q.e.d.)
3
Exercise 5.4
(i) Let f(t, x) = ln x. We have
f
t
= 0,
f
x
=
1
x
,

2
f
x
2
=
1
x
2
and
(dS(t))
2
=
2
(t)S
2
(t)dt.
Applying Itos lemma, the dierential of the log stock price d ln S(t) becomes
d ln S(t) = df(t, S(t))
=
1
S(t)
dS(t)
1
2
1
S
2
(t)
(dS(t))
2
= r(t)dt + (t)d

W(t)
1
2

2
(t)dt
=
_
r(t)
1
2

2
(t)
_
dt + (t)d

W(t).
In integral form, we get
ln S(t) = ln S(0) +
_
t
0
_
r(s)
1
2

2
(s)
_
ds +
_
t
0
(s)d

W(s). (2)
Taking the exponential yields
S(t) = S(0)e
X(t)
where
X(t) =
_
t
0
_
r(s)
1
2

2
(s)
_
ds +
_
t
0
(s)d

W(s).
By Theorem 4.4.9, Ito integrals of a deterministic integrand are normally distributed
with zero mean. Thus, X(t) is normally distributed with
X
t
N
__
t
0
_
r(s)
1
2

2
(s)
_
ds,
_
t
0

2
(s)ds
_
(3)
4
(ii) Since the payo of a European call option only depends on the asset price S
T
at
maturity, two dierent diusion processes for the underlying that have imply a
common risk-neutral density for S
T
at time T will yield the same option prices. By
setting r(t) = R and (t) = in Equation (3), we see that the distribution of X
t
under constant interest rate and volatility is given by
X
t
N
__
t
0
_
R
1
2

2
_
ds,
_
t
0

2
ds
_
N
__
R
1
2

2
_
T,
2
T
_
. (4)
We now equate the mean and variance in Equations (3) and (4) to obtain

2
T =
_
t
0

2
(s)ds =
1
T

_
t
0

2
(s)ds
and
_
R
1
2

2
_
T =
_
t
0
_
r(s)
1
2

2
(s)
_
ds R =
_
t
0
r(s)ds
It follows that we can replace R by
_
t
0
r(s)ds and by
_
_
t
0

2
(s)ds in the Black-
Scholes formula for European call options to obtain the value under deterministic
interest rates and volatilities.
Exercise 5.5
(i) Let f(t, x) =
1
x
. We have
f
t
= 0,
f
x
=
1
x
2
,

2
f
x
2
=
2
x
3
and thus
d
_
1
Z(t)
_
=
1
Z
2
(t)
dZ(t) +
1
2
2
Z
3
(t)
(dZ(t))
2
=
(t)
Z(t)
dW(t) +

2
Z(t)
dt.
5
Here, we have used that
dZ(t) = (t)Z(t)dW(t),
as shown in the proof of Theorem 5.2.3.
(ii) As shown in the proof of Theorem 5.2.3, Z(t) is a Radon-Nikod ym derivative process.
Thus, Lemma 5.2.2 applies and we get
E
_

M(t)Z(t)

F(s)
_
= Z(s)

E
_

M(t)

F(s)
_
= Z(s)M(s) (q.e.d.).
Here, the rst equality follows by multiplying both sides of the equality in Lemma
5.2.2 by Z(s) and in the second step we use that

M(t) is a martingale under

P.
(iii) By the Ito product rule (Corollary 4.6.3), we get
d

M(t) = d
_
M(t)
Z(t)
_
= M(t)d
_
1
Z(t)
_
+
1
Z(t)
dM(t) + dM(t)d
_
1
Z(t)
_
=
M(t)
Z(t)
_

2
(t)dt + (t)dW(t)

+
(t)
Z(t)
dW(t) +
(t)(t)
Z(t)
dt
=
1
Z(t)
__
M(t)
2
(t) + (t)(t)
_
dt + (M(t)(t) + (t)) dW(t)

.
(iv) Corollary 5.3.2 denes
d

W(t) = dW(t) + (t)dt
and substituting for W(t) in the dierential of

M(t) yields
d

M(t) =
1
Z(t)
_
_
M(t)
2
(t) + (t)(t)
_
dt + (M(t)(t) + (t))
_
d

W(t) (t)dt
__
=
1
Z(t)
(M(t)(t) + (t)) d

W(t)
=
_

M(t)(t) +
(t)
Z(t)
_
d

W(t).
6
We now dene

(t) =

M(t)(t) +
(t)
Z(t)
and integrate to obtain

M(t) =

M(0) +
_
t
0

(u)d

W(u) (q.e.d.).
Note, that (t) is adapted to the ltration F(t) as required by Corollary 5.3.2, since

M(t), (t), (t) and Z(t) are adapted processes.

Exercise 5.6
We will proof the more general case of the multidimensional Girsanov theorem, i.e. The-
orem 5.4.2 for any d {1, 2, . . .}. We note, that by the two dimensional Levy theorem
(Theorem 4.6.5) two continuous martingales M
1
(t) and M
2
(t) with respect to some ltra-
tion F(t) that start at zero and have unit quadratic variation and zero cross variation are
independent Brownian motions. If we thus consider the multidimensional martingale
M(t) = (M
1
(t), M
2
(t), . . . , M
d
(t)) ,
then we require that the conditions of the two dimensional Levy theorem are satised
for any two M
i
(t) and M
j
(t) where i, j {1, 2, . . . , d}. I.e. we want to show that
dM
i
(t)dM
j
(t) =
_

_
dt if i = j
0 otherwise
or equivalently in matrix notation
dM(t)dM

(t) = I
d
dt,
where

is used to denote the vector/matrix transpose and I
d
is the identity matrix of
dimension d.
(i) Continuity:
7
Clearly,

W(t) = W(t) +
_
t
0
(u)du
is continuous since by Denition 3.3.1 the Brownian motion W(t) has continuous
sample paths. Similarly, the ordinary Lebesgue integral is a continuous function of
the upper limit of integration.
(ii) Starting at zero:
It is obvious that

W(0) = W(0) = 0.
(iii) Unit quadratic and zero cross variation:
We have
d

W(t)d

W

(t) = (dW(t) + (t)dt) (dW(t) + (t)dt)

= (dW(t) + (t)dt) (dW

(t) +

(t)dt)
= dW(t)dW

(t) = I
d
dt (q.e.d.).
(iv) Martingale property:
We rst dene
X(t) =
_
t
0
(u) dW(u)
1
2
_
t
0
||(u)||
2
du
such that
dX(t) = (t) dW(t)
1
2
||(t)||
2
dt
and
8
dX(t)dX(t) =
_

j=1

j
(t)dW
j
(t)
1
2
d

j=1

2
j
(t)dt
_
2
=
d

j=1
d

k=1

j
(t)
k
(t)dW
j
(t)dW
k
(t)
=
d

j=1

2
j
(t)dt
= ||(t)||
2
dt.
In analogy to the proof of Theorem 5.2.3 (Girsanov, one dimension), we now dene
f(t, x) = e
x
and apply the It o formula to obtain the dierential of Z(t) as
dZ(t) = Z(t)dX(t) +
1
2
Z(t)dX(t)dX(t)
= Z(t)(t) dW(t)
1
2
Z(t)||(t)||
2
dt +
1
2
Z(t)||(t)||
2
dt
= Z(t)(t) dW(t).
Since the dierential of Z(t) contains no dt term, it follows that Z(t) is a P martingale
with E[Z(T)] = Z(0) = 1 and thus qualies as a Radon-Niko ym derivative process.
The dierential of

W(t)Z(t) can be computed using the It o product rule (Corollary
4.6.3) as
d
_

W(t)Z(t)
_
=

W(t)dZ(t) + Z(t)d

W(t) + d

W(t)dZ(t)
=

W(t)Z(t)(t) dW(t) + Z(t) (dW(t) + (t)dt)

+(dW(t) + (t)dt) (Z(t)(t) dW(t))
=

W(t)Z(t)(t) dW(t) + Z(t)dW(t) + Z(t)(t)dt

Z(t)(t)dt
=

W(t)Z(t)(t) dW(t) + Z(t)dW(t)

= Z(t)
_

W(t) (t) + 1
_
dW(t).
By the same argument at before, it follows that

W(t)Z(t) is a P martingale. Using
Lemma 5.2.2, we obtain
9

E
_

W(t)

F(s)
_
=
1
Z(s)
E
_

W(t)Z(t)

F(s)
_
=
1
Z(s)

W(s)Z(s) =

W(s) (q.e.d.).
Thus,

W(t) is a d-dimensional

P martingale.
Since all conditions of the multidimensional Levy theorem are satised, we conclude
that

W(t) is a d-dimensional Brownian motion under

P.
Exercise 5.7
(i) We can obtain this portfolio process by investing the amount X
2
(0) > 0 in the
money market such that the value at time T is
X
2
(0) exp
__
T
0
R(t)dt
_
=
X
2
(0)
D(T)
.
In addition, we invest in the portfolio process X
1
(t). The initial cost of setting up
this portfolio are X
2
(0) + X
1
(0) = X
2
(0). At time T, we have
P
_
X
2
(T)
X
2
(0)
D(T)
_
= P
_
X
2
0
D(T)
+ X
1
(T)
X
2
(0)
D(T)
_
= P{X
1
(T) 0} = 1
and
P
_
X
2
(T) >
X
2
(0)
D(T)
_
= P
_
X
2
0
D(T)
+ X
1
(T) >
X
2
(0)
D(T)
_
= P{X
1
(T) > 0} > 0.
(ii) The same idea as in (i) applies here. We can obtain the portfolio process X
1
(t) by
borrowing the amount X
2
(0) in the money market and investing the proceeds in the
portfolio process X
2
(t). At time T, we have to repay X
2
(0)/D(T) and obtain
P{X
1
(T) 0} = P
_
X
2
(T)
X
2
(0)
D(T)
0
_
= P
_
X
2
(T)
X
2
(0)
D(T)
_
= 1
and
P{X
1
(T) > 0} = P
_
X
2
(T)
X
2
(0)
D(T)
> 0
_
= P
_
X
2
(T) >
X
2
(0)
D(T)
_
> 0.
10
Exercise 5.8 (Every strictly positive Asset is a generalized geo-
metric Brownian Motion)
(i) We have that
D(t)V (t) =

E[ D(T)V (T)| F(t)] .
is a

P-martingale since for s < t

E[ D(t)V (t)| F(s)] =

E
_

E[ D(T)V (T)| F(t)]

F(s)
_
=

E[ D(T)V (T)| F(s)]
= D(s)V (s).
By Corollary 5.3.2, there exists a an adapted process

(t) such that
D(t)V (t) = V (0) +
_
t
0

(u)d

W(u)
or in dierential form
d(D(t)V (t)) =

(t)d

W(t).
By the It o product rule,
dV (t) = D
1
(t)d(D(t)V (t)) + D(t)V (t)dD
1
(t) + d(D(t)V (t))dD
1
(t)
= R(t)V (t)dt +

(t)
D(t)
d

W(t).
Here we used that
dD
1
(t) = R(t)D
1
(t)dt
is a process of nite variation and thus the cross variation term in the It o dierential
drops out.
11
(ii) We dene a new random variable X = 0 with E[X] = 0. Note that the constant
X satises Denition 1.2.1 for random variables. Since both D(T) and V (T) are

P-almost surely positive, it follows that X < D(T)V (T)

P-almost surely. Conse-
quently, 0 = E[X] < E[D(T)V (T)] = D(t)V (t). This does not follow directly from
Theorem 1.3.1 but is a straight forward modication of it. Now, since D(t) is positive

P-almost surely, it follows that V (t) is too.

(iii) Since V (t) is

P-almost surely positive, we can write its dierential as
dV (t) = R(t)V (t)dt +

(t)
D(t)V (t)
V (t)d

W(t)
= R(t)V (t)dt + (t)V (t)d

W(t),
where we dened (t) =

(t)/(D(t)V (t)). (t) is adapted to the ltration F(t) since
all three processes that dene it are.d
Exercise 5.9 (Implying the risk-neutral distribution)
We apply the Leibniz integral rule to obtain
c
K
=

K
_
e
rT
_

K
(y K) p(0, T, x, y)dy
_
= e
rT
_

K

K
(y K) p(0, T, x, y)dy
= e
rT
_

K
p(0, T, x, y)dy.
Applying the Leibniz integral rule for a second time yields
c
KK
= e
rT
p(0, T, x, K).
Thus, the implied risk-neutral density can be recovered from a continuum of call prices
by
p(0, T, x, K) = e
rT
c
KK
.
12
Exercise 5.10 (Chooser Option)
i) At time t
0
(choice date) the holder of a long position in a chooser option has the right
to choose whether his option is a call or a put. If he acts rationally, he will always
choose the higher priced option. Thus, the price V (t
0
) of a chooser option at t = t
0
has to be the greater of the call option price C (t
0
) and the put option price P (t
0
).
Using put-call parity we obtain
V (t
0
) = max {C (t
0
) , P (t
0
)}
= max {C (t
0
) , C (t
0
) F (t
0
)}
= C (t
0
) + max {0, F (t
0
)}
= C (t
0
) +
_
e
r(Tt
0
)
K S (t
0
)
_
+
. (q.e.d.)
ii) By the risk-neutral pricing formula we have
V (0) = E
Q
_
e
rt
0
V (t
0
)

= E
Q
_
e
rt
0
_
C (t
0
) +
_
e
r(Tt
0
)
K S (t
0
)
_
+
__
= E
Q
_
e
rt
0
E
Q
_
e
r(Tt
0
)
C(T)

F (t
0
)

+E
Q
_
e
rt
0
_
e
r(Tt
0
)
K S (t
0
)
_
+
_
= C(0) +E
Q
_
e
rt
0
max
_
0, e
r(Tt
0
)
K S (t
0
)
_
= C(0, T, K) + P(0, t
0
, e
r(Tt
0
)
K). (q.e.d)
While the rst term is the current value of a call option with strike price K and
maturity in T, the second term is the current value of a put option with strike price
e
r(Tt
0
)
K and maturity in t
0
. Thus the value of a chooser option is equal to a
portfolio consisting of a long call and a long put with dierent strike prices and
expiry dates.
Exercise 5.11 (Hedging a Cash Flow)
Following the hint, we start by dening the market price of risk by
(t) =
(t) R(t)
(t)
.
13
The process (t) is adapted to the ltration F(t) since all three processes that dene
it are. It is well dened since (t) is strictly positive. Next, we dene
Z(t) = exp
_

_
t
0
(u)dW(u)
1
2
_
t
0

2
(u)du
_
and

W(t) = W(t) +
_
t
0
(u)du.
By Girsanovs theorem (Theorem 5.2.3),

W(t) is a Brownian motion under the prob-
ability measure

P dened by

P(A) =
_
A
Z()dP()
for all A F. Further following the hint, we dene

M(t) =

E
__
T
0
D(u)C(u)du

F(t)
_
.
Then, for 0 s t we get

E
_

M

F(s)
_
=

E
_

E
__
T
0
D(u)C(u)du

F(t)
_

F(s)
_
=

E
__
T
0
D(u)C(u)du

F(s)
_
=

M(s)
and thus

M(t) is a

P-martingale. Since the ltration F(t) is the one generated by the
Brownian motion W(t), we can apply Corollary 5.3.2. There exists an adapted process

(t) such that

M(t) =

M(0) +
_
t
0

(u)d

W(u).
Using It os product rule, we now compute the dierential of the discounted portfolio
value under P as
14
d(D(t)X(t)) = D(t)dX(t) + X(t)dD(t) + dD(t)dX(t)
= D(t)(t)dS(t) + R(t)D(t)(X(t) Delta(t)S(t))dt D(t)C(t)dt R(t)D(t)X(t)
= (t)D(t)(dS(t) R(t)S(t)dt) D(t)C(t)dt
= (t)D(t)S(t)(((t) R(t))dt + (t)dW(t)) D(t)C(t)dt
= (t)D(t)S(t)
_
((t) R(t))dt + (t)
_

W(t) (t)dt
__
D(t)C(t)dt
= (t)(t)D(t)S(t)d

W(t) D(t)C(t)dt.
In integral form, we have
D(T)X(T) = X(0) +
_
T
0
(u)(u)D(u)S(u)d

W(u)
_
T
0
D(u)C(u)du.
where we used that D(0) = 1. Choose X(0) =

M(0) and
(t) =

(t)
(t)D(t)S(t)
,
then
D(T)X(T) =

M(0) +
_
T
0

(u)d

W(0)
_
T
0
D(u)C(u)du
=

M(0) +

M(T)

M(0)
_
T
0
D(u)C(u)du
=

E
__
T
0
D(u)C(u)du

F(T)
_

_
T
0
D(u)C(u)du
= 0.
Here, we used that the integral in the second last equality is F(T)-measurable. Since
D(T) is

P-almost surely positive, it follows that X(T) = 0

P-almost surely. It follows
that if we start with an initial wealth of
X(0) =

E
__
T
0
D(u)C(u)
_
and choose the portfolio process given by (t) above, then we can hedge the random
cash-ow

P-almost surely.
15
Exercise 5.12 (Correlation under Change of Measure)
(i) Remember that B
i
(t) is dened by
B
i
(t) =
d

j=1
_
t
0

ij
(u)

i
(u)
dW
j
(u).
We have

B
i
(t) =
d

i=1
_
t
0
_

ij
(u)

i
(u)
dW
j
(u) +
i
(u)du
_
=
d

i=1
_
t
0

ij
(u)

i
(u)
(dW
j
(u) + (u)du)
=
d

i=1
_
t
0

ij
(u)

i
(u)
d

W
j
(u).
It has been shown using Levys theorem (Theorem 4.6.4) that B
i
(t) is a Brownian
motion under P. Repeating the same proof but replacing W
j
(t) by

W
j
(t) it can be
shown that

B
i
(t) is a Brownian motion under P. Note that by Theorem 4.3.1, each
of the It o integrals in the denition of

B
i
(t) is a continuous martingale starting at
zero and thus

B
i
(t) is too. Its quadratic variation is
d

B
i
(t)d

B
i
(t) =
d

j=1

2
ij
(t)

2
i
(t)
dt
= dt,
where we used that by denition

i
(t) =

_
d

j=1

2
ij
(t).
As anticipated, it follows by Theorem 4.6.4 that

B
i
(t) is a

P-martingale.
(ii) We have
16
dS
i
(t) =
i
(t)S
i
(t)dt +
i
(t)S
i
(t)dB
i
(t)
=
i
(t)S
i
(t)dt +
i
(t)S
i
(t)
d

j=1

ij
(t)

i
(t)
dW
j
(t)
=
i
(t)S
i
(t)dt +
i
(t)S
i
(t)
d

j=1

ij
(t)

i
(t)
_
d

W
j
(t)
j
(t)dt
_
=
i
(t)S
i
(t)dt +
i
(t)S
i
(t)
_
d

B
i
(t)
d

j=1

ij
(t)
j
(t)

i
(t)
dt
_
=
_

i
(t)
d

j=1

ij
(t)
j
(t)
_
S
i
(t)dt +
i
(t)S
i
(t)d

B
t
= R(t)S
i
(t)dt +
i
(t)S
i
(t)d

B
i
(t) (q.e.d.).
Here, we used the expression previously obtained for

B
i
(t) in equality four. The last
equality follows from the market price of risk equations

i
(t) R(t) =
d

j=1

ij
(t)
j
(t).
(iii) We have for i = k
d

B
i
(t)d

B
k
(t) = (dB
i
(t) +
i
(t)dt) (dB
k
(t) +
k
(t)dt)
= dB
i
(t)dB
k
(t)
=
ik
(t)dt (q.e.d.).
Here, we used that the covariation between a process of bounded variation and any
other process is zero in the second step.
(iv) Using the Ito product rule, we get for the rst expectation
17
E[B
i
(t)B
k
(t)] =

E
_
B
i
(0)B
k
(0) +
_
t
0
d (B
i
(u)B
k
(u))
_
=

E
__
t
0
(B
i
(u)dB
k
(u) + B
k
(u)dB
i
(u) + dB
i
(u)dB
k
(u))
_
=

E
__
t
0
dB
i
(u)dB
k
(u)
_
=

E
__
t
0

ik
(u)du
_
=
_
t
0

ik
(u)du.
Here, we used that B
i
(0) = B
k
(0) = 0 in the second equality. By Theorem 4.3.1,
the Ito integrals in the third equality are martingales starting at zero and have thus
zero expected value. In the last step, we can drop the expectation operator since
the function
ik
(t) is deterministic. The same steps can be repeated to show that

E
_

B
i
(t)

B
k
(t)
_
=
_
t
0

ik
(u)du
and the proof is complete.
(v) Again using the Ito product rule, we get for the rst expectation
E[B
1
(t)B
2
(t)] =

E
_
B
2
(0)B
2
(0) +
_
t
0
d (B
1
(u)B
2
(u))
_
=

E
__
t
0
(B
1
(u)dB
2
(u) + B
2
(u)dB
1
(u) + dB
1
(u)dB
2
(u))
_
=

E
__
t
0
dB
1
(u)dB
2
(u)
_
=

E
__
t
0
sign (W
1
(u)) du
_
=
_
t
0
E[sign (W
1
(u))] du
=
_
t
0
(P{W
1
(u) 0} P{W
1
(u) < 0}) du
= 0.
Skipping some intermediate steps, we get for the second expectation
18

E
_

B
1
(t)

B
2
(t)
_
=
_
t
0

E[sign (W
1
(u))] du
=
_
t
0

E
_
sign
_

W
1
(u)
_
u
0

1
(v)dv
__
du
=
_
t
0

E
_
sign
_

W
1
(u) u
__
du
=
_
t
0
_

P
_

W
1
(u) u
_

P
_

W
1
(u) < u
__
du
0.
Here, strict inequality holds for all t > 0. Consequently, for t > 0,
E[B
1
(t)B
2
(t)] =

E
_

B
1
(t)

B
2
(t)
_
(q.e.d.).
Exercise 5.13
(i) We have

E[W
1
(t)] =

E
_

W
1
(t)
_
t
0

1
(u)du
_
= 0,
where the last equality follows from

W
1
being a

P-martingale with zero initial value
and
1
(t) = 0 for all t 0. Similarly,

E[W
2
(t)] =

E
_

W
2
(t)
_
t
0

2
(u)du
_
=

E
_

_
t
0
W
1
(u)du
_
=

E
_

_
t
0
_

W
1
(u)
1
(u)
_
du
_
=

E
_

_
t
0

W
1
(u)du
_
.
To evaluate the remaining expectation, we dene a function f(t, x) = tx where
f
t
= x,
f
x
= t,

2
f
x
2
= 0.
19
Applying Itos formula yields the dierential
d
_
t

W
1
(t)
_
= df
_
t,

W
1
(t)
_
=

W
1
(t)dt + td

W
1
(t).
Thus, in integral form and after rearranging, we get

_
t
0

W
1
(u)du =
_
t
0
ud

W
1
(u) t

W
1
(t).
Substituting back yields

E[W
2
(t)] =

E
__
t
0
ud

W
1
(u) t

W
1
(t)
_
= 0 (q.e.d.).
Here, we used that by Theorem 4.3.1, both the It o integral w.r.t.

W(t) and

W(t)
have zero expectation under

P.
(ii) We have

Cov [W
1
(T), W
2
(T)] =

E[W
1
(T)W
2
(T)]

E[W
1
(T)]

E[W
2
(T)]
=

E[W
1
(T)W
2
(T)] .
Here, the second equality follows from the result in (i). Now using the hint, we have
in integral form
W
1
(T)W
2
(T) =
_
T
0
W
1
(u)dW
2
(u) +
_
T
0
W
2
(u)dW
1
(u)
=
_
T
0
W
1
(u)
_
d

W
2
(u) W
1
(u)du
_
+
_
T
0
W
2
(u)d

W
1
(u)
=
_
T
0
W
1
(u)d

W
2
(u)
_
T
0

W
2
1
(u)du +
_
T
0
W
2
(u)d

W
1
(u).
Thus,
20

Cov [W
1
(T), W
2
(T)] =

E
__
T
0
W
1
(u)d

W
2
(u)
_
T
0

W
2
1
(u)du +
_
T
0
W
2
(u)d

W
1
(u)
_
=

E
_

_
T
0

W
2
1
(u)du
_
=
_
T
0

E
_

W
2
1
(u)
_
du
=
_
T
0
udu
=
1
2
T
2
q.e.d..
Here, we again used the martingale property of the two Ito integrals w.r.t.

W
1
(t)
and

W
2
(t) under

P in the second equality.
Exercise 5.14 (Cost of Carry)
(i) The dierential of the discounted portfolio value is given by
d(D(t)X(t)) = D(t)dX(t) + X(t)dD(t) + dD(t)dX(t)
= (t)D(t)dS(t) a(t)D(t)dt + rD(t)(X(t) (t)S(t))dt rD(t)X(t)
= (t)D(t) (dS(t) rS(t)dt) a(t)D(t)dt
= (t)D(t)S(t)d

W(t).
Since

W(t) is a

P-martingale, the claim follows.
(ii) Let f(t, x) = e
x
where
f
t
= 0,
f
x
= e
x
,

2
f
x
2
= e
x
and dene
X(t) =
_
r
1
2

2
_
t +

W(t).
Then,
21
dY (t) = df(t, X(t))
= Y (t)dX(t) +
1
2
Y (t)dX(t)dX(t)
=
_
r
1
2

2
_
Y (t)dt + Y (t)d

W(t) +
1
2

2
dt
= rY (t)dt + Y (t)d

W(t) (q.e.d.).
Furthermore,
d(D(t)Y (t)) = D(t)dY (t) + Y (t)dD(t) + dD(t)dY (t)
= rD(t)Y (t)dt + D(t)Y (t)d

W(t) rD(t)Y (t)dt
= D(t)Y (t)d

W(t) (q.e.d.).
Since the dierential contains no dt-term, it follows that D(t)Y (t) is a

P-martingale.
Finally, from the second denition of S(t), we obtain the dierential form by applying
the Ito product rule
dS(t) = S(0)dY (t) + Y (t)d
__
t
0
a
Y (s)
ds
_
+
__
t
0
a
Y (s)
ds
_
dY (t) + dY (t)d
__
t
0
a
Y (s)
ds
_
=
_
S(0) +
_
t
0
a
Y (s)
ds
_
dY (t) + Y (t)
a
Y (t)
dt
= r
_
S(0) +
_
t
0
a
Y (s)
ds
_
Y (t)dt +
_
S(0) +
_
t
0
a
Y (s)
ds
_
Y (t)d

W(t) + adt
= rS(t)dt + S(t)d

W(t) + adt (q.e.d.).
We could have directly obtained the solution to the linear stochastic dierential
equation for S(t) using the (yet to be obtained) result from Exercise 6.1. Using the
notation of that exercise, we have
dS(t) = (a(t) + b(t)X(t))du + ((t) + (t)S(t))d

W(t)
where
a(t) = a, b(t) = r, (t) = 0, (t) = .
22
Let
A(t) = exp
__
t
0
(v)d

W(v) +
_
t
0
_
b(v)
1
2

2
(v)dv
__
= exp
_

W(t) +
_
r
1
2

2
_
t
_
and
B(t) = S(0) +
_
t
0
a(v) (v)(v)
A(v)
dv +
_
t
0
(v)
Z(v)
d

W(v)
= S(0) +
_
t
0
a
A(v)
dv.
Then the solution to the stochastic dierential equation is
S(t) = A(t)B(t)
= S(0)A(t) + A(t)
_
t
0
a
A(v)
dv.
This coincides with the result that we just proved.
(iii) Following the hint, we get

E[ S(T)| F(t)] =

E
_
S(0)Y (T) + Y (T)
_
T
0
a
Y (s)
ds

F(t)
_
= S(0)

E[ Y (T)| F(t)] +

E[ Y (T)| F(t)]
_
t
0
a
Y (s)
ds
+

E
_
Y (T)
_
T
t
a
Y (s)
ds

F(t)
_
= S(0)

E[ Y (T)| F(t)] +

E[ Y (T)| F(t)]
_
t
0
a
Y (s)
ds
+a
_
T
t

E
_
Y (T)
Y (s)

F(t)
_
ds.
We further have
23

E[ Y (T)| F(t)] =

E
_
exp
__
r
1
2

2
_
T +

W(t)
__
=

E
_
Y (t) exp
__
r
1
2

2
_
(T t) +
_

W(T)

W(t)
_
_

F(t)
_
= Y (t)e
r(Tt)

E
_
exp
_

1
2

2
(T t) +
_

W(T)

W(t)
_
__
= Y (t)e
r(Tt)

E
_
exp
_

1
2

2
(T t) +

W(T t)
__
= Y (t)e
r(Tt)
.
The third equality uses that Y (t) is F(t)-measurable and that the increment

W(T)

W(t) is independent of the -algebra F(t) to drop the conditioning. In the fourth
equality we exploit the time homogeniety of the Brownian motion, i.e. that the law
of

W(T)

W(t) under

P is the same as that of

W(T t). Finally, we use that
e
rt
Y (t) = exp
_

1
2

2
t

W(t)
_
is a

P-martingale with an initial value of one as shown in (ii). Next, for t s T
and using similar steps we have

E
_
Y (T)
Y (s)

F(t)
_
=

E
_
exp
__
r
1
2

2
_
(T s) +
_

W(T)

W(s)
_
__
= e
r(Ts)

E
_
exp
_

1
2

2
(T s) +

W(T s)
__
= e
r(Ts)
.
Thus,

E[ S(T)| F(t)] = S(0)Y (t)e

r(Tt)
+ Y (t)e
r(Tt)
_
t
0
a
Y (s)
ds + a
_
T
t
e
r(Ts)
ds
= S(t)e
r(Tt)

a
r
e
r(Ts)

s=T
s=t
= S(t)e
r(Tt)

a
r
_
1 e
r(Tt)
_
.
(iv) Let f(t, x) = xe
r(Tt)

a
r
_
1 e
r(Tt)
_
where
f
t
= rxe
r(Tt)
ae
r(Tt)
,
f
x
= e
r(Tt)
,

2
f
x
2
= 0.
24
An application of It os lemma yields the dierential
d
_
S(t)e
r(Tt)

a
r
_
1 e
r(Tt)
_
_
=
_
rS(t)e
r(Tt)
ae
r(Tt
_
dt + rS(t)e
r(Tt)
dt + S(t)e
r(Tt)
d

W(t) + ae
r(Tt)
dt
= S(t)e
r(Tt)
d

W(t).
Since the dierential contains no dt-term it follows that

E[ S(T)| F(t)] is a martingale
under

P. This is also obvious by letting s t T and directly computing

E
_
S(t)e
r(Tt)

a
r
_
1 e
r(Tt)
_

F(s)
_
=

E
_

E[ S(T)| F(t)]

F(s)
_
=

E[ S(T)| F(s)]
= S(s)e
r(Ts)

a
r
_
1 e
r(Ts)
_
(q.e.d.).
(v) This result immediately follows from the denition of the value of the forward con-
tract. We have

E
_
e
r(Tt)
(S(T) For
S
(t, T))

F(t)

For
S
(t, T) =

E[ S(T)| F(t)] (q.e.d.).
Here, we used that For
S
(t, T) is F(t)-measurable to take it outside the conditional
expectation.
(vi) Following the hint, we rst compute the dierential of the discounted portfolio value
d(D(t)X(t)) = D(t)dX(t) + X(t)dD(t) + dD(t)dX(t)
= (t)D(t)dS(t) a(t)D(t)dt + rD(t)(X(t) (t)S(t))dt rD(t)X(t)dt
= (t)D(t)(dS(t) rS(t)dt) a(t)D(t)dt
= (t)d(D(t)S(t)) a(t)D(t)dt.
Integrating and using that (t) = 1 for all 0 t T yields
25
D(T)X(T) =
_
T
0
d(D(t)S(t)) a
_
T
0
D(t)dt
= D(t)S(t)|
t=T
t=0
a
_
T
0
e
rt
dt
= D(T)S(T) S(0) +
a
r
_
e
rT
1
_
.
Solving or X(T) yields
X(T) = S(T) e
rT
_
S(0)
a
r
_
e
rT
1
_
_
= S(T)
_
S(0)
a
r
_
1 e
rT
_
_
= S(T) For
S
(0, T) (q.e.d.).
26