Seminars on Continuous Time Finance

Raquel M. Gaspar

Fall 2003

1
Contents

3 Stochastic Integrals 4

4 Differential Equations 7

6 Arbitrage Pricing 10

7 Completeness and Hedging 13

8 Parity Relations and Delta Hedging 14

9 Several Underlying Assets 17

10 Incomplete Markets 19

11 Dividends 20

12 Currency Derivatives 21

15 Bonds and Interest Rates 24

16 Short Rate Models 26

17 Martingale Models for the Short Rate 28

18 Forward Rate Models 32

19 Change of Numeraire 34

20 Extra Exercises 38

20.1 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

20.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

21 Exams 49

2
21.1 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

21.1.1 March 26, 2003 . . . . . . . . . . . . . . . . . . . . . . . . 49

21.1.2 January 10, 10.00-15.00, 2003 . . . . . . . . . . . . . . . . 51

21.1.3 March 18, 2002 . . . . . . . . . . . . . . . . . . . . . . . . 53

21.2 Topics of solutions . . . . . . . . . . . . . . . . . . . . . . . . . . 55

21.2.1 March 26, 2003 . . . . . . . . . . . . . . . . . . . . . . . . 55

21.2.2 January 10, 2003 . . . . . . . . . . . . . . . . . . . . . . . 58

21.2.3 March 18, 2002 . . . . . . . . . . . . . . . . . . . . . . . . 60

3
3 Stochastic Integrals

Exercise 3.1

(a) Since Z(t) is determinist, we have

dZ(t) = αeαt dt
= αZ(t)dt.

(b) By definition of a stochastic differential

dZ(t) = g(t)dW (t)

(c) Using Itô’s formula

α2 αW (t)
dZ(t) = e dt + αeαW (t) dW (t)
2
α2
= Z(t)dt + αZ(t)dW
2
(d) Using Itô’s formula and considering the dynamics of X(t) we have
α2 αx
dZ(t) = αeαx dX(t) + e (dX(t))2
 2 
1 2 2
= Z(t) αµ + α σ dt + ασZ(t)dW (t).
2

(e) Using Itô’s formula and considering the dynamics of X(t) we have

dZ(t) = 2X(t)dX(t) + (d(X(t))2

 
= Z(t) 2α + σ 2 dt + 2ZσdW (t).

Exercise 3.3 By definition we have that the dynamics of X(t) are given by
dX(t) = σ(t)dW (t).

Consider Z(t) = eiuX(t) . Then using the Itô’s formula we have that the dynamic
of Z(t) can be described by
 2 
u
dZ(t) = − σ 2 (t) Z(t)dt + [iuσ(t)] Z(t)dW (t)
2

From Z(0) = 1 we get,

Z t Z t
u2 2
Z(t) = 1 − σ (s)Z(s)ds + iu σ(s)Z(s)dW (s).
2 0 0

4
Taking expectations we have,
Z t  Z t 
u2 2
E [Z(t)] = 1 − E σ (s)Z(s)ds + iuE σ(s)Z(s)dW (s)
2 0 0
Z t 
u2
= 1− σ 2 (s)E [Z(s)] ds + 0
2 0

By setting E [Z(t)] = m(t) and differentiating with respect to t we find an

ordinary differential equation,
∂m(t) u2
= − m(t)σ 2 (t)
∂t 2
with the initial condition m(0) = 1 and whose solution is
 2Z t 
u
m(t) = exp − σ2(s)ds
2 0
= E [Z(t)]
h i
= E eiuX(t)

So, X(t) is normally distributed. By the properties of the normal distribution

the following relation
h i u2
E eiuX(t) = eiuE[X(t)]− 2 V [X(t)]

where V [X(t)] is the variance of X(t), so it must be that E [X(t)] = 0 and

Rt
V [X(t)] = 0 σ 2 (s)ds.

Exercise 3.5 We have a sub martingale if E [ X(t)| Fs ] ≥ X(s)∀, t ≥ s. From

the dynamics of X we can write
Z t Z t
X(t) = X(s) + µ(z)dz + σ(z)dW (z).
s s

By taking expectation, conditioned at time s, from both sides we get

Z t  

E [ X(t)| Fs ] = E [ X(s)| Fs ] + E µ(z)dz  Fs
s
  


Z t  
  
= X(s) + E s  µ(z)dz  Fs 
 s  
| {z }
≥0

≥ X(s)

so X is a sub martingale.

5
Exercise 3.6 Set X(t) = h(W1 (t), · · · , Wn (t)).

We have by Itô that

Xn n
∂h 1 X ∂2h
dX(t) = dWi (t) + dWi (t)dWj (t)
i=1
∂xi 2 i,j=1 ∂xi ∂xj

∂h ∂2h
where ∂xi denotes the first derivative with respect to the i-th variable, ∂xi ∂xj
denotes the second order cross-derivative between the i-th and j-th variable and
all derivatives should be evaluated at (W1 (s), · · · , Wn (s)).

Since we are dealing with independent Wiener processes we know

∀u : dWi (u)dWj (u) = 0 for i 6= j and dWi (u)dWj (u) = du for i = j,

so, integrating we get

Z tXn Z t n
X
∂h 1 ∂2h
X(t) = dWi (u) + dWi (u)dWj (u)
0 i=1 ∂x i 2 0 i,j=1
∂xi ∂xj
Z tXn Z n
∂h 1 tX ∂2h 2
= dWi (u) + [dWi (u)]
0 i=1 ∂x i 2 0 i=1 ∂xi ∂xj
Z tXn Z t Xn
∂h 1 ∂2h
= dWi (u) + du.
0 i=1 ∂x i 2 0 i,j=1
∂xi ∂xj

Taking expectations
"Z  #   
n
tX  Z t Xn 
∂h  1 ∂ 2
h 
E [ X(t)| Fs ] = E dWi (u) Fs + E  du Fs 
0 i=1 ∂xi  2 0 i,j=1 ∂xi ∂xj 
Z n
sX Z s n
X
∂h 1 ∂2h
= dWi (u) + du
0 i=1
∂xi 2 0 i,j=1 ∂xi ∂xj
| {z }
X(s)

"Z  #   
n
tX  Z t Xn 
∂h  1 2
∂ h 
+E dWi (u) Fs +E  du Fs 
0 i=1 ∂xi  2 s i,j=1 ∂xi ∂xj 
| {z }
0
  
Z t Xn 
1 2
∂ h 
= X(s) + E  du Fs  .
2 s i,j=1 ∂xi ∂xj 

Pn ∂2h
• If h is harmonic the last term is zero, since i,j=1 ∂xi ∂xj = 0, we have

E [ X(t)| Fs ] = X(s) so X is a martingale.

6
 Pn ∂2h
• If h is subharmonic the last term is always nonnegative, since i,j=1 ∂xi ∂xj ≥
0 we have

E [ X(t)| Fs ] ≥ X(s) so X is a submartingale.

Exercise 3.8

(a) Using the Itô’s formula we find the dynamics of R(t),

1 
dR(t) = 2X(t)(dX(t)) + 2Y (t)(dY (t)) + 2(dX(t))2 + 2(dY (t))2
  2
= (2α + 1) X 2 (t) + Y 2 (t) dt
= (2α + 1)R(t)dt

From the dynamics we can see immediately that R(t) is deterministic (it
has no stochastic component!).

(b) Integrating the SDE for X(t) and taking expectations we have
Z t
X(t) = x0 + α E [X(s)] ds
0

Which once more can be solve setting m(t) = E [X(t)],taking the deriva-
tive with respect to t and using ODE methods, to get the answer

E [X(t)] = x0 eαt

4 Differential Equations

Exercise 4.1 We have:

dY (t) = αeαt x0 dt, dZ(t) = αeαt σdt, dR(t) = e−αt dW (t).

Itô’s formula then gives us (the cross term dZ(t) · dR(t) vanishes)

dX(t) = dY (t) + Z(t) · dR(t) + R(t) · dZ(t)

Z t
αt αt −αt
= αe x0 dt + e · σ · e dW (t) + e−αs dW (s) · αeαt σdt
0
 Z t 
αt α(t−s)
= α e x0 + σ e dW (s) dt + σdW (t)
0
= αX(t)dt + σdW (t).

7
Exercise 4.5 Using the dynamics of X(t) and the Itô formula we get
 
1 2
dY (t) = αβ + β(β − 1)σ Y (t)dt + σβY (t)dW (t)
2
= µY (t)dt + δY (t)dW (t)

where µ = αβ + 12 β(β − 1)σ 2 and δ = σβ so Y is also a GBM.

Exercise 4.6 From the Itô formula and using the dynamics of X and Y
1 X(t) 1 X(t)
dZ(t) = dX(t) − dY (t) − dX(t)dY (t) + (dY (t))2
Y (t) Y (t)2 Y (t)2 Y (t)3
 
= Z(t) α − γ + δ 2 dt + σZ(t)dW (t) − δZ(t)dV (t).

Exercise 4.9 From Feyman-Kac we have We have

F (t, x) = E t,x [2 ln[X(T )]] ,

and

dX(s) = µX(s)ds + σXdW (s),

X(t) = x.

Solving the SDE, we obtain (check the solution of the GBM in th extra exercises
if you do not remmeber)
 
1 2
X(T ) = exp ln x + (µ − σ )(T − t) + σ[W (T ) − W (t)] ,
2
and thus
1
F (t, x) = 2 ln(x) + 2(µ − σ 2 )(T − t).
2

Exercise 4.10 Given the dynamics of X(t) any F (t, x) that solves the problem
has the dynamics given by

∂F ∂F 1 ∂2F
dF (t, x) = dt + dX(t) + (dX(t))2
∂t ∂x 2 ∂x2
∂F ∂F
= dt + [µ(t, x)dt + σ(t, x)dW (t)] + k(t, x)dt − k(t, x)dt
∂t ∂x

8
 1 ∂2F  2 
+ 2
σ (t, x)dW (t)
2 ∂x 

 

 ∂F ∂F 1 2 
= + µ(t, x) + σ (t, x) + k(t, x) dt − k(t, x)dt

 ∂t ∂x {z2 }

| 
0
∂F
+ σ(t, x)dW (t)
∂x
∂F
= −k(t, x)dt + σ(t, x)dW (t)
∂x
We now write F (T, X(T )) in terms of F (t, x) and the dynamics of F during the
time period t . . . T (recall that we defined X(t) = x)
Z T Z T
∂F
F (t, X(T )) = F (t, x) − k(s, X(s)ds + σ(s, X(s))dW (s)
t t ∂x
⇔
Z T Z T
∂F
F (t, x) = F (T, X(T )) + k(s, X(s)ds − σ(s, X(s))dW (s)
t t ∂x

Taking expectations Et,x [.] from both sides

"Z #
T
F (t, x) = Et,x [F (T, X(T ))] + Et,x k(s, X(s)ds
t
Z T
= Et,x [Φ(T )] + Et,x [k(s, X(s)] ds
t

Exercise 4.11 Using the representation formula from Exercise 4.10 we get
Z T
 
F (t, x) = Et,x 2 ln[X 2 (T )] + Et,x [X(s)] ds,
t

Given
dX(s) = X(s)dW (s).

The first term is easily computed as in the exercise 4.9 above. Furthermore it
is easily seen directly from the SDE (how?)that Et,x [X(s)] = x. Thus we have
the result

F (t, x) = 2 ln(x) − (T − t) + x(T − t)

= ln(x2 ) + (x − 1)(T − t)

9
6 Arbitrage Pricing

Exercise 6.1

(a) From standard theory we have

Π (t) = F (t, S(t)), where F solves the Black-Scholes equation.
Using Itô we obtain
 
∂F ∂F 1 ∂2F ∂F
dΠ (t) = + rS(t) + σ 2 S 2 (t) 2 dt + σS(t) dW (t).
∂t ∂s 2 ∂s ∂s
Using the fact that F satisfies the Black-Scholes equation, and that F (t, S(t)) =
Π (t) we obtain
∂F
dΠ (t) = rΠ (t) dt + σS(t) dW (t)
∂s
and so g(t) = σS(t) ∂F
∂s .
Π(t)
(b) Apply Itô’s formula to the process Z(t) = B(t) and use the result in (a).

1 Π(t)
dZ(t) = (dΠ(t)) − 2 (d(B(t))
B(t) B (t)
g(t)
= dW (t)
B(t)
σS(t) ∂F
= Z(t) dW (t)
Π(t) ∂s
Z is a martingale since Et [Z(T )] = Z(t) for all t < T and its diffusion
coefficient is given by σZ (t) = σS(t) ∂F
Π(t) ∂s .

Exercise 6.4 We have as usual

Q  β 
Π (t) = e−r(T −t) Et,s S (T ) .

We know from earlier exercises (check exercises 3.4 and 4.5) that Y (t) = S β (t)
satisfies the SDE under Q
 
1 2
dY (t) = rβ + β(β − 1)σ Y (t)dt + σβY (t)dW (t).
2
Using the standard technique, we can integrate, take expectations, differentiate
with respect to time and solve by ODE techniques, to obtain
Q  β  1 2
Et,s S (T ) = sβ e[rβ+ 2 β(β−1)σ ](T −t) ,

10
So,
1 2
Π(t) = sβ e[r(β−1)+ 2 β(β−1)σ ](T −t) .

Exercise 6.6 We consider only the case when t < T0h. The iother case is handled
Q S(T1 )
in very much the same way. We have to compute Et,s S(T0 ) . Define the process
X on the time interval [T0 , T1 ] by
S(u)
X(u) = .
S(T0 )
Q
We now want to compute Et,s [X(T1 )]. The stochastic differential (under Q) of
X is easily seen to be

dX(u) = rXdu + σXdW (u),

X(T0 ) = 1.

From this SDE it follows at once (the same technique of integrating, taking
expectations, differentiate with respect to time and solve by ODE techniques)
that
Q
Et,s [X(T1 )] = er(T1 −T0 ) ,
and thus the price, at t of the contract is given by

Π (t) = e−r(T0 −t) .

Exercise 6.7 The price in SEK of the ACME INC., Z, is defined as Z(t) =
S(t)Y (t) and by Itô has the following dynamics under Q

dZ(t) = rZ(t)dt + σZ(t)dW1 (t) + δZ(t)dW2 (t)

We also have, by using Itô once more, that the dynamics of ln Z 2 are
 
d ln Z 2 (t) = 2r − σ 2 − δ 2 dt + 2σdW1 (t) + 2δdW2 (t)

which integrating and taking conditioned expectations give us

Q    
Et,z ln[Z 2 (T )] = ln z 2 + 2r − σ 2 − δ 2 (T − t)

Since we know that

Q  
Π(t) = F (t, s) = e−r(T −t) Et,z ln[Z 2 (T )] ,

11
the arbitrage free pricing function Π is
  
Π(t) = e−r(T −t) ln z 2 + 2r − σ 2 − δ 2 (T − t)
  
= e−r(T −t) 2 ln(sy) + 2r − σ 2 − δ 2 (T − t) ,

where, as usual, z = Z(t), s = S(t) and y = Y (t).

Exercise 6.9 The forward price, i.e. the amount of money to be payed out at
time T , but decided at the time t is

F (t, T ) = EtQ [X ] .

Note that the forward price is not the price of the forward contract on the
T -claim X which is what we are looking for.

Take for instance the long position: at time T , the buyer of a forward contract
receives X and pays F (t, T ). Hence, the price at time t of that position is
  
 −r(T −t)  
Π(t; X − F (t, T )) = EtQ 
e
X − F (t, T ) = 0.
 | {z }
EtQ [X ]

At time s > t, however, the underlying asset may have changed in value, in a
way different from expectations, so then the price of a forward contract can be
defined as
h i
Π(s; X − F (t, T )) = EsQ e−r(T −s) (X − F (t, T ))
 
F (t,T )
 Q z }| {
= e−r(T −s)  Q
Es [X ] − Et [X ] .


Remark: For the special case where the contract is on one share S we get:
 
 Q 
Π(s) = e−r(T −s) 
Es [S(T )] − S(t)e
r(T −t) 
.
| {z }
EtQ [S(T )]

We can also easily calculate EsQ [S(T )] since

Z s Z T
EsQ [S(T )] = S(t) + r S(u)du +r EsQ [S(u)] du
| {zt } s

S(s)

12
so,
EsQ [S(T )] = S(s)er(T −s)
and, therefore, the free arbitrage pricing function at time s > t is

Π(s) = S(s) − S(t)er(s−t) .

7 Completeness and Hedging

Exercise 7.2 We have F (t, s, z) be defined by

1
Ft + r · s · Fs + σ 2 s2 Fss + gFz = rF
2
F (T, s, z) = Φ(s, z)

and the dynamics under Q for S and Z

dS(u) = rS(u)du + σS(u)dW (u)

dZ(u) = g(u, S(u))du
Q
We want to show that F (t, S(t), Z(t)) = e−r(T −t) Et,s,z [Φ(S(T ), Z(T ))].

For that we find , by Itô, the dynamics of Π(t) = F (t, S(t), Z(t)), the arbitrage
free pricing process
1
dΠ(t) = Ft dt + Fs [(rS(t)dt + σS(t)dW (t)] + Fz · g(t, S(t))dt + Fss σ 2 S 2 (t)dt
  2
1
= Ft + r · S(t) · Fs + σ 2 S 2 (t)Fss + g(t, S(t))Fz +σS(t)Fs dW (t)
2
| {z }
rΠ(t)

Integrating we have
Z T Z T
Π(T ) = Π(t) + r Π(u)du + σ S(u)Fs dW (u)
t t

Hence Z T
Q Q
Et,z,s [Π(T )] = Π(t) + r Et,z,s [Π(u)] du
t
Q
So, using the usual ”trick” of setting m(u) = Et,z,s [Π(u)] and using techniques
of ODE we finally get
Q
Π(t) = F (t, S(t), Z(t)) = e−r(T −t) Et,s,z [Φ(S(T ), Z(T ))] .

(Remember that Π(T ) = F (T, S(T ), Z(T )) = Φ(S(T ), Z(T )).)

13
Exercise 7.3 The price arbitrage free price is given by (note that this time our
claim is not simple, i.e. it is not of the form X = Φ(S(T ))).

Π(t) = e−r(T2 −t) EtQ [X ]

Z T2
1
= e −r(T2 −t)
EtQ [S(u)] du
T2 − T1 T1

We know that under Q

dS(u) = rS(u)du + σS(u)dW (u)

S(t) = s

So,
⇒ EtQ [S(u)] = ser(u−t)
Z T2
1 1 s h r(T2 −t) i
ser(u−t) du = e − er(T1 −t)
T 2 − T 1 T1 T2 − T1 r
The price to the ”mean” contract is thus
s h i
Π(t) = 1 − e−r(T2 −T1 ) .
r(T2 − T1 )

8 Parity Relations and Delta Hedging

Exercise 8.1 The T -claim X given by:


 K, if S(T ) ≤ A
X = K + A − S(T ), if A < S(T ) < K + A ,

0, otherwise.
has then following contract function (recall that X = ΦS(T ))

 K, if x ≤ A
Φ(x) = K + A − x, if A < x < K + A ,

0, otherwise.
which can be decomposed into the following ”basic” contract functions written

Φ(x) = K · |{z}
1 − max [0, x − A] + max [0, x − A − K] .
| {z } | {z }
ΦB (x) Φc,A (x) Φc,A+K (x)

Having this T-claim X is then equivalent to having the following (replicating)

portfolio at time T :

14
 * K in monetary units

* short (position in) a call with strike A

* long (position in) a call with strike A + K

Given the decomposition of the contract function Φ into basic contract functions,
we immediately have that the arbitrage free pricing process Π is
B(t)
z }| {
Π(t) = K · e−r(T −t) −c(s, A, T ) + c(s, A + K, T )

where c(s, A, T ) and c(s, A + K, T ) stand for the prices of European call options
on S and maturity T with strike prices A and A + K, respectively. The Black-
Scholes formula give us both c(s, A, T ) and c(s, A + K, T ) .

The hedge portfolio thus consists of a reverse position in the above components,
i.e., borrow e−r(T −t) K, buy a call with strike K and sell a call with strike A+K.

Exercise 8.4 We apply, once again, the exact same technique. The T -claim X
given by: 
 0, if S(T ) < A


 S(T ) − A, if A ≤ S(T ) ≤ B
X =

 C − S(T ), if B < S(T ) ≤ C


0, if S(T ) > C.
A+C
where B = 2 , has a contract function Φ that can be written as

Φ(x) = max [0, x − A] + max [0, x − C] −2 max [0, x − B]

| {z } | {z } | {z }
Φc,A (x) Φc,C (x) Φc,B (x)

Having this butterfly is then equivalent to having the following constant(replicating)

portfolio at time T :

* long (position in) a call option with strike A

* long (position in) a call option with strike C

* short (position in) a call option with strike B

15
The arbitrage free pricing process Π follows immediately from the decomposition
of the contract function Φ and is given by

Π(t) = c(s, A, T ) + c(s, C, T ) − 2c(s, B, T )

where c(s, A, T ), c(s, B, T ) and c(s, C, T ) stand for the prices of European call
options on S, with maturity T and strike prices A, B and C, respectively, and
can be computed using the Black-Scholes formula.

The hedge portfolio consists of a reverse position in the above components, i.e.,
sell two call options one with strike A and other with strike B and buy other
two both with strike B.

Exercise 8.10 From the put-call parity we have that

p(t, s) = Ke−r(T −t) + c(t, s) − S

where p(t, s) and c(t, s) stand for the price of a put and a call option on S with
maturity T and strike price K.

The delta measures the variation in the price of a derivative with respect to
changes in the value of the underlying. Differentiating the put-call parity w.r.t.
S we have
∂p(t, s) ∂  −r(T −t) ∂c(t, s) ∂S
= Ke + −
| ∂S
{z } ∂S | ∂S
{z } ∂S
∆put ∆call
∆put = ∆call − 1

Since,∆call = N [d1] ⇒ ∆put = N [d1] − 1.

To find the result on the gamma we differentiate one more time (so two times)
w.r.t. S the put-call parity and get

∂ 2 p(t, s) ∂ 2 c(t, s)
2
= 2
| ∂S {z } | ∂S {z }
Γcall Γput

ϕ(d1) ϕ(d1)
From the fact that Γcall = √
sσ T −t
it follows that Γput = √
sσ T −t
.

16
9 Several Underlying Assets

Exercise 9.3

Consider the T -claim X = max [S1 (T ) − S2 (T ); 0] where S1 and S2 are defined

by
dS1 (t) = α1 S1 (t)dt + σ1 S1 (t)dW̄1

dS2 (t) = α2 S2 (t)dt + σ2 S2 (t)dW̄2

and
dW̄1 .dW̄2 = ρdt.

The contract function Φ associated with it, is homogeneous of degree 1, and

thus, it can be rewritten as a contract function Ψ on the normalized variable
s1
z= s2 .
 
 s1 
Φ(s1 , s2 ) = max [s1 − s2 ; 0] = s2 max  
 s2 −1; 0
|{z}
z
= s2 max [z − 1; 0] .
| {z }
Ψ(z)

So, the pricing function F (t, s1 , s2 ) = s2 G(t, z) and from

Ft = s2 Gt
∂z 1
F1 = s2 Gz = s2 Gz = Gz
∂s1 s2
 
∂z s1
F2 = G + s2 Gz = G + s2 Gz − 2 = G − zGz
∂s2 s2
∂z 1
F11 = Gzz = Gzz
∂s1 s
2 
∂z ∂z ∂z 1
F22 = Gz − − Gz + zGzz = z 2 Gzz
∂s2 ∂s2 ∂s2 s2
∂z 1
F12 = Gzz = − zGzz
∂s2 s2
we realize that G satisfies the (much easier) PDE,
1
Gt + z 2 Gzz (σ12 + σ22 − 2σ1 σ2 ρ) = 0
2
G(T, z) = max [z − 1, 0] ,

17
that we recognize as the Black-Scholes equation and, therefore we have that G
is the price of a call with underlying Z, exercise price K = 1 and with, r = 0,
p
and σ = σ12 + σ22 − 2ρσ1 σ2 .

So, we can use Black-Scholes formula to concretize the pricing function Π(t) =
F (t, S1 (t), S2 (t)):

Π(t) = S2 (t)G(t, Z(t))

 
Z(t)

 z }| { 


 S (t) 

1 0 0
= S2 (t) N [d1 ] − N [d2 ]

 S2 (t) 


 


= S1 (t)N [d01 ] − S2 (t)N [d02 ]

where
   
1 S1 (t) 1 2

d01
=p 2 √ ln + σ1 + σ22 − 2ρσ1 σ2 (T − t)
σ1 + σ22 − 2ρσ1 σ2 T − t S2 (t) 2
p √
and d02 = d01 − σ12 + σ22 − 2ρσ1 σ2 T − t.

Exercise 9.4 For this ”special” maximum option we have X = max [aS1 (T ); bS2 (T )]

where S1 and S2 are defined by

dS1 (t) = α1 S1 (t)dt + σ1 S1 (t)dW̄1

dS2 (t) = α2 S2 (t)dt + σ2 S2 (t)dW̄2

and W̄1 is independent from W̄2 .

The contract function Φ is homogeneous of degree 1 and can be rewritten in

terms of the normalized variable z = ss12 :
 
as1
Φ(s1 , s2 ) = max [as1 ; bs2 ] = bs2 max ;1
bs2
   
as1
= = bs2 max − 1; 0 + 1
bs2
  z  

a z}|{ 

 s1 b 
= bs2 max  − ; 0 + 1

b s2 a 

 
b
= bs2 + as2 max z − ; 0
a
| {z }
Ψ(z)

18
Thus, the pricing function F is given by F (t, s1 , s2 ) = bs2 + as2 G(t, z) where G
solves the boundary problem So we have
1
Gt + z 2 Gzz (σ12 + σ22 ) = 0
2  
b
G(T, z) = max z − , 0 .
a

Once again, the above expression is the Black-Scholes equation, G is the price of
p
a call option on Z with strike price K = ab , r = 0, and σ = σ12 + σ22 We can use
Black-Scholes formula to concretize the pricing function Π(t) = F (t, S1 (t), S2 (t):

Π(t) = bS2 (t) + aS2 (t)G(t, Z(t))
 
S1 (t) b
= bS2 (t) + aS2 (t) N [d01 ] − N [d02 ]
S2 (t) a
= bS2 (t) + aS1 (t)N [d01 ] − bS2 (t)N [d02 ]
= (1 − N [d02 ])bS2 (t) + aN [d01 ] S1 (t)

where
   
1 aS1 (t) 1 2

d01 = p 2 √ ln + σ + σ 2
(T − t)
σ1 + σ22 T − t bS2 (t) 2 1 2

p √
and d02 = d01 − σ12 + σ22 T − t.

10 Incomplete Markets

Exercise 10.1 Given a claim X = Π(X(T ))) and since the dynamics of X
under the Q-measure are

dX(t) = [µ(t, X(t)) − λ(t, X(t))σ(t, X(t))] dt + σ(t, X(t))dW Q (t),

we can find the Q-dynamics of the pricing function F (t, X(t)) using the Ito
formula
∂F ∂F 1 ∂2F
dF (t, X(t)) = dt + dX(t) + (dX(t))2
∂t ∂x 2 ∂x2
∂F ∂F
= dt + [µ(t, X(t)) − λ(t, X(t))σ(t, X(t))] dt + σ(t, X(t))dW Q (t)
∂t ∂x
1 ∂2F 2
+ σ (t, X(t))dt
2 ∂x2

19
  
∂F ∂F 1 ∂2F 2
= + (µ(t, X(t)) − λ(t, X(t))σ(t, X(t))) + σ (t, X(t)) dt
∂t ∂x 2 ∂x2
| {z }
rF (t,X(t))
Q
+σ(t, X(t))dW (t)
= rF (t, X(t)) + σ(t, X(t))dW Q (t)

where the last step results from the fact that F has to satisfy the pricing PDE:

∂F ∂F 1 ∂2F 2
+ (µ(t, X(t)) − λ(t, X(t))σ(t, X(t))) + σ (t, X(t)) = rF.
∂t ∂x 2 ∂x2

11 Dividends

Exercise 11.2 We know that when there is a continuous dividend δ being paid
we have the following dynamics under Q for the asset S and the dividen structure
D

dS(t) = [r − δ (S(t))] S(t)dt + σ (S(t)) S(t)dW (t)

dD(t) = S(t)δ (S(t)) dt

Just by rewritting and rearranging terms we have

dS(t) = rS(t)dt − δ (S(t)) S(t)dt +σ (S(t)) S(t)dW (t)

| {z }
dD(t)
−r·t −r·t −r·t
e dS(t) − e rS(t)dt = −e dD(t) + e−r·t σ (S(t)) S(t)dW (t)
 −r·t 
d e S(t) = −e−r·t dD(t) + e−r·t σ (S(t)) S(t)dW (t).

Integratting the last expression we get

Z t Z t
−r·t −r·0 −ru
e S(t) = e S(0) − e dD(u) + e−r·u σ (S(u))S(u)dW (u)
0 0
Z t Z t
−r·t −ru
S(0) = e S(t) + e dD(u) − e−r·u σ (S(u))S(u)dW (u),
0 0

and taking E0Q [.] expectations we finally get the results

 Z t 
S(0) = E0Q e−r·t S(t) + e−ru dD(u) .
0

20
Exercise 11.6 In the Black-Scholes model with a constant continuous dividend
yield δ we have, under the Q-measure we have

dS(t) = (r − δ) S(t)dt + σS(t)dW Q (t).

From the “standard” call-put parity, which must be valid, we have the following
relation between call and put options with the same maturity T and exercise
price K:
p(t, x) = c(t, x) − Π(t, S(T )) + e−r(T −t) K.

But we also know that

Π(t, S(T )) = e−r(T −t) EtQ [S(T )]

= e−r(T −t) S(t)e(r−δ)(T −t)
= S(t)e−δ(T −t) .

So,
p(t, x) = c(t, x) − S(t)e−δ(T −t) + e−r(T −t) K.

12 Currency Derivatives

Exercise 12.1 We have, under the objective probability measure, the following
processes for the spot exchange rate X (units of domestic currency d per foreign
currency unit f ), and the domestic , Bd , and foreign, Bf , riskless assets:

dX(t) = αx X(t)dt + σx X(t)dW (t)

dBd (t) = rd B(t)d dt
dBf (t) = rf B(t)d dt,

where rd and rf are the domestic and foreign short rates which are assumed to
be deterministic. Hence the Q-dynamics of X are given by

dX(t) = (rd − rf )X(t)dt + σx X(t)dW (t).

From the “standard” call-put parity, which must be valid, we have the following
relation between call and put options with the same maturity T and exercise
price K:
p(t, x) = c(t, x) − Π(t, X(T )) + e−rd (T −t) K.

21
But we also know that

Π(t, X(T )) = e−rd (T −t) EtQ [X(T )]

= e−rd (T −t) X(t)e(rd −rf )(T −t)
= X(t)e−rf (T −t) .

So,
p(t, x) = c(t, x) − X(t)e−rf (T −t) + e−rd (T −t) K.

(Remark: Compare this exercise with exercise 11.6 in the previous section, and
see that the foreign risk-free rate can be treated as a continuous dividend-yield
on the spot exchange rate.)

Exercise 12.2 The binary option on the exchange rate X is a T -claim, Z, of

the form
Z = 1[a,b] (X(T )) ,

i.e. if a ≤ X(T ) ≤ b then the holder of this claim will obtain one unit of domestic
currency, otherwise gets nothing. The dynamics under Q of the exchange rate
X are given by

dX(t) = (rd − rf )X(t)dt + σx X(t)dW (t).

where rd and rf are the domestic and foreign short rates which are assumed to
be deterministic.

Integrating the above expression we get

Z T Z t
X(T ) = X(t) + (rd − rf )X(u)du + σx X(u)dW (u)
t t

and so we have (by solving the SDE)

 2

σx
rd −rf − 2 (T −t)+σx (W (T )−W (t))
X(T ) = X(t)e ,

and we see from taking the logarithm that

   
σ2 √
ln(X(T )) ∼ N ln (X(t)) + rd − rf − x (T − t), σx T − t .
2

22
So, the price, Π(t) of the binary option is given by

Π(t) = e−rd (T −t) EtQ [Z]

= e−rd (T −t) Q (a ≤ X(T ) ≤ b)
= e−rd (T −t) Q (ln (a) ≤ ln(X(T )) ≤ ln (b))
= e−rd (T −t) Q (da ≤ z ≤ db )


a
2

σx b

 2
σx

ln X(t)
− rd −rf − 2 (T −t) ln X(t)
− rd −rf − 2 (T −t)
where z ∼ N [0, 1] and da = √ and db =
σx T −t
√
σx T −t
and so we have that the price of the binary exchange option Z is:

Π(t) = e−rd (T −t) [N (db ) − N (da )] .

Exercise 12.3 Under the objective probability measure we have the following
dynamics for the domestic stock Sd , the foreign stock Sf , the exchange rate X
and the domestic Bd and foreign Bf riskless assets

dSd (t) = αd Sd (t)dt + σd Sd (t)dWd (t),

dSf (t) = αf Sf (t)dt + σf Sf (t)dWf (t),
dX(t) = αx X(t)dt + σx X(t)dW (t),
dBd (t) = rd Bd (t)dt,
dBf (t) = rf Sd (t)dt,

where rd and rf are the domestic and foreign short rates which are assumed
to be deterministic. The domestic stock denominated in terms of the foreign
currency is given by S˜d = Sd , then by Ito
X

1 Sd (t) Sd (t) −1
dS˜d (t) = dSd (t) − 2 dX(t) + 3 (dX(t))2 − 2 dSd (t)dX(t)
X(t) X (t) X (t) X (t) | {z }
0 for Wd and W indep.

= αd S˜d (t)dt + σd S˜d (t)dWd (t) − αx S˜d (t)dt − σx S˜d (t)dW (t) + σx2 S˜d (t)dt


= αd − αx + σx2 S˜d (t)dt + σd S˜d (t)dWd (t) − σx S˜d (t)dW (t).

The dynamics above are under the objective probability measure. Under Qf
the drift term of all assets denominated in the foreign currency must have rf
(the risk-free rate on the foreign economy) as the drift. Also, we can use the
properties of the Wiener processes to normalize the diffusion part.

23
It follows that
q
dS˜d (t) = rf S˜d (t)dt + σd2 + σx2 S˜d (t)dWf (t).

15 Bonds and Interest Rates

Exercise 15.1 Forward Rate Agreement

(a) Note that the cash flow to the lender’s in a FRA (−K at time S and
∗
KeR (T −S)
at time T ) can be replicated by the following portfolio:
* sell K S-bonds
∗
* buy KeR (T −S)
T -bonds
So, at time t < S, the value Π(t), on the lender’s cash flow in a FRA has
to equal to the value of the replicating portfolio and is given buy
∗
Π(t) = KeR (T −S)
p(t, T ) − Kp(t, S).

(b) If we have that Π(0) = 0 we must have

∗
KeR (T −S)
p(0, T ) − Kp(0, S) = 0
R∗ (T −S)
e p(0, T ) − p(0, S) = 0
 
p(0, S)
R∗ (T − S) = ln
p(0, T )
ln(p(0, S)) − ln(p(0, T ))
R∗ =
T −S

Since by definition the forward rate R(t; S, T ) is given by

ln(p(t, T )) − ln(p(t, S))

R(t; S, T ) = − ,
T −S
we have that R∗ = R(0; S, T ).

Exercise 15.2 Since f (t, T ) = − ∂ ln(p(t,T

∂T
))
and p(t, T ) satisfies

dp(t, T ) = p(t, T )m(t, T )dt + p(t, T )v(t, T )dW (t),

24
then we have, by Ito, that ln(p(t, T )) has dynamics that are given by
1
d ln(p(t, T )) = m(t, T )dt + v(t, T )dW (t) − v(t, T )v(t, T )∗ dt
 2
1 ∗
= m(t, T ) − v(t, T )v(t, T ) dt + v(t, T )dW (t),
2

and integrating we have

Z t Z t
1
ln(p(t, T )) = ln(p(0, T )) + m(s, T ) − v(s, T )v(s, T )∗ ds + v(s, T )dW (s).
0 2 0

∂ ln(p(t,T ))
Since f (t, T ) = − ∂T , we also have
 
Z Z t
∂ ln(p(0, T )) t  ∂m(s, T ) ∂v(s, T )  ∂v(s, T )
f (t, T ) = − −  −v(s, T )  ds− dW (s)
∂T  ∂T 
| {z } 0 | {z } | ∂T
{z } 0 ∂T
f (0,T ) mT (s,T ) vT (s,T )∗

and from its differential form we finally get the result

 
 
df (t, T ) = v(t, T )vT (t, T )∗ − mT (t, T ) dt − vT (t, T ) dW (t).
| {z } | {z }
α(t,T ) σ(t,T )

Exercise 15.5 Let {y(0, T ; T ≥ 0} denote the zero coupon yield curve.

(a) Then we have that

p(0, T ) = e−y(0,T )·T .

Using the definition of the instantaneous forward rate f (t, T ) = − ∂ ln(p(t,T

∂T
))

and the expression above we get the result

∂ ln (p(0, T ))
f (0, T ) = −
∂T

∂ ln e−y(0,T )·T
= −
∂T
∂ (y(0, T ) · T )
=
∂T
∂y(0, T )
= y(0, T ) + T · .
∂T

25
 (b) If the zero coupon yield curve is an increasing function of T , then we know
that ∂y(0,T
∂T
)
≥ 0, and using the result from (a) we have that f (0, T ) ≥
y(0, T ) for any T > 0.
It remains to prove that yM (0, T ) ≤ y(0, T ). This will follow from the
price of a coupon bond is given by
n
X
pT (t) = Kp(t, Tn ) + ci p(t, Ti ).
i=1

So in particular we have that

n
X
pT (0) = K e|−y(0,T
{z
n )·Tn
}+ ci e|−y(0,T i )·Ti
{z }.
p(0,Tn ) i=1 p(0,Ti )

but it can also we given, since yM is the yield to maturity of a coupon

bond, by
n
X
pT (0) = Ke−yM (0,Tn )·Tn + ci e−yM (0,Tn )·Ti .
i=1

So,
n
X n
X
K e|−y(0,T
{z
n )·Tn
}+ ci e|−y(0,T i )·Ti
{z } = Ke
−yM (0,Tn )·Tn
+ ci e−yM (0,Tn )·Ti
p(0,Tn ) i=1 p(0,Ti ) i=1

By comparing the LHS and the RHS and since ci ≥ 0, and y(0, Tn ) ≥
y(0, Tn−1 ) ≥ · · · ≥ y(0, T1 ) (by assumption), we must have that y(0, Tn ) ≥
yM (0, Tn ) since it is valid for any Tn we can write that y(0, T ) ≥ yM (0, T )
for any T .

16 Short Rate Models

Exercise 16.2 The object of the exercise is to connect the forward rates to the
risk neutral valuation of bond prices.

(a) Recall from the risk-neutral valuation of bond prices that

" ( Z )#
T
Q
p(t, T ) = Et exp − r(s)ds
t

26
 and hence, using the definition of a instantaneous forward rate (and as-
suming that we can differentiate under the expectation sign) we get
∂ ln (p(t, T ))
f (t, T ) = −
∂T " ( Z )#!
T
∂ Q
= − ln Et exp − r(s)ds
∂T t
h n R oi
T
EtQ r(T ) exp − t r(s)ds
= h n R oi .
T
EtQ exp − t r(s)ds

(b) To check that indeed we have f (t, t) = r(t) use the expression above
h n R oi
T
EtQ r(T ) exp − t r(s)ds
f (t, T ) = h n R oi ,
T
EtQ exp − t r(s)ds

and set T = t:
h n R oi
t
EtQ r(t) exp − t r(s)ds
f (t, t) = h n R oi
t
EtQ exp − t r(s)ds

EtQ [r(t) exp {0}]

= = r(t).
EtQ [exp {0}]

Exercise 16.3 Recall that in a swap of a fixed rate vs. a short rate we have:

• A invests K at time 0 and let it grow at a fixed rate of interest R over

the time interval [0, T ]. A has thus an amount KA at time T and at that
time (T ) pays the surplus KA − K to B.

• B invests the principal at a stochastic short rate of interest over the interval
[0, T ]. B has thus an amount KB at time T and at that time (T ) pays the
surplus KB − K to A.

• The swap rate for this contract is defined as the value, R, of the fixed rate
which gives this contract value zero at t = 0.
RT
r(s)ds
At maturity party A has Φ(T ) = KB − KeR·T where KB = Ke 0 . Thus,
the value of this contract at time 0 is given by
  
RT RT
Q  − r(s)ds  r(s)ds 
Π(0) = E0,r e 0 Ke 0 − KeR·T 
| {z }
Φ(T )

27
  RT 
Q R·T − r(s)ds
= E0,r K − Ke 0

 RT 
Q − r(s)ds
= K · E0,r 1 − eR·T e 0
 
  RT 
 Q − r(s)ds 
= K · 1 − eR·T E0,r e 0 
 
| {z }
p(0,T )

Since we must have Π(0) = 0 we have

K · 1 − eR·T p(t, T ) = 0
1
eR·T =
p(0, T )
R·T = − ln (p(0, T ))
ln (p(0, T ))
R = − .
T

17 Martingale Models for the Short Rate

Exercise 17.1 In the Vasicek model we have

dr(t) = (b − ar(t))dt + σdW (t)

with a > 0.

(a) Using the SDE above and multiplying both sides by eat we have that

eat dr(t) + r(t)aeat dt = eat bdt + σeat dW (t)

d(eat r(t)) = eat bdt + σeat dW (t)
Z t Z t
eat r(t) = r(0) + eas bds + σ eas dW (s)
0 0
Z t
b 
r(t) = r(0)e−at + 1 − e−at + σe−at ea (s)dW (s).
a 0

Looking at the solution of the SDE it is immediate that r is Gaussian, so

it is enough to determine the mean and variance.
From the solution of the SDE we see that
b 
E [r(t)] = r(0)e−at + 1 − e−at
a

28
 and
Z t
V [r(t)] = σ 2 e−2at e2as ds
0
σ2

= 1 − e−2at .
2a
(b) As t → ∞ we have
bh i b
lim E [r(t)] = r(0) lim e−at + 1 − lim e−at = ,
t→∞ t→∞ a t→∞ a
and
σ2   σ2
lim V [r(t)] = 1 − lim e−2at = .
t→∞ 2a t→∞ 2a
h i
b σ2
So, N a , 2a is the limiting distribution of r.

(c) The results in (a) and (b) are based on the assumption
h i that the value r(0)2
is known. If, instead we have that r(0) ∼ N ab , σ2a , then for any t we
have that
b  b
E [r(t)] = E [r(0)] e−at + 1 − e−at = ,
a a
and
σ2

V [r(t)] = e−2at V [r(0)] + 1 − e−2at
2a
σ 2 e−2at σ2
σ2
= + 1 − e−2at = .
2a 2a 2a
(d) To be done!

Exercise 17.2 From Exercise 17.1 we know that for the Vasicek model, and
t < u we have
Z u
bh i
r(u) = r(t)e−a(u−t) + 1 − e−a(u−t) + σ e−a(u−s) dW (s).
a t

By definition,  RT 
Q − r(u)du
p(t, T ) = Et,r e t .
RT
Let us define Z(t, T ) = − t r(u)du, from exercise 17.1 we know that r is
normally distributed, so Z is also normally distributed since
Z T " Z T # Z TZ u
at −au b at −au
Z(t, T ) = −r(t)e e du − 1−e e du + σ e−a(u−s) dW (s)du
t a t t t
Z TZ u
r(t) h −a(T −t)
i b b2 h −a(T −t)
i
= − 1−e − (T − t) − 1−e +σ e−a(u−s) dW (s)du
a a a t t

29
and has
 i
Q 1h −a(T −t) b b2 h i
Et,r [Z(t, T )] = − 1−e r(t) − (T − t) − 1 − e−a(T −t)
a | a a{z }
| {z }
deterministic function of t and T deterministic function of t and T

and
" Z Z #
u T
Q −a(u−s)
Vt,r [Z(t, T )] = V σ e dudW (s)
t t
Z Z !2
u T
−a(u−s)
= σ e du ds .
t t
| {z }
deterministic function of t and T

Since Z(t, T ) is normally distributed ∀t,T (from the properties of the normal
distribution), we have that
h i Q 1 Q
Q
p(t, T ) = Et,r eZ(t,T ) = eEt,r [Z(t,T )]+ 2 Vt,r [Z(t,T )]
⇒
Q 1 Q
ln(p(t, T )) = Et,r [Z(t, T )] + Vt,r [Z(t, T )]
2
1 h i
= − 1 − e−a(T −t) r(t) +
a
| {z }
B(t,T )
 
 Z u Z !2 
 b 2 h i T 
+ − (T − t) − b 1 − e−a(T −t) + 1 σ −a(u−s)
e du ds
 a a 2 t .
 t 
| {z }
A(t,T )

So, the Vasicek model has an affine term structure.

Alternative Solution

From Exercise 17.1 we know that for the Vasicek model, and t < u we have
Z u
bh i
r(u) = r(t) e|−a(u−t)
{z } a+ 1 − e−a(u−t)
+ σ e−a(u−s) dW (s) .
D(u) | {z t }
F (u)

where D(u) is deterministic and F (u) is stochastic and both are non dependent
on r.

By definition,
 RT 
Q − r(u)du
p(t, T ) = Et,r e t

30
  RT 
Q − [r(t)D(u)+F (u)du]
= Et,r e t
 
 R
 R  
 − tT D(u)du r(t) Q − tT F (u)du 
= |e {z E
} t,r e .
 
B | {z }
A
RT
In part A, since − t F (u)du is normally distributed (note it is a “sum” of
Wiener increments), we know (from the properties of the normal distribution)
that there exist a A(t, T ) such that
 RT 
Q − F (u)du
Et,r e t = eA(t,T ) .
RT
From part B we immediatelly see that B(t, T ) = t D(u)du.

So, the Vasicek model has an affine term structure.

Exercise 17.3 The problem with the Dothan’s model is that when r follows a
GBM, like
dr(t) = ar(t)dt + σr(t)dW (t),
the solution to the SDE is, for any t < u,
1 2
r(u) = r(t)e(a− 2 σ )(u−t)+σ(W (u)−W (t)) ,

i.e., in the Dothan’s model r is lognormally distributed. Since,

 RT 
Q − r(u)du
p(t, T ) = Et,r e t ,

to use the same procedure as in exercise 17.2 we would have to compute the
above expected value, i.e., the expected value of an integral (that is a “sum”)
of lognormally distributed variables, which is a mess!

Exercise 17.8 Take the follwoing CIR model

p
dY (t) = 2aY (t) + σ 2 dt + 2σ Y (t)dW (t), Y (0) = y0 ,
p
Then by Ito Z(t) = Y (t) follows
 
1 1
 1 1 3 2
dZ(t) = Y (t)− 2 dY (t) + − Y (t)− 2 (dY (t))
2 2 4
1 1 1
 1 1
= 2aY (t) 2 + σ 2 Y (t)− 2 dt + σdW (t) − σ 2 Y (t)− 2
2 2
1
= a Y (t) 2 dt + σdW (t),
| {z }
Z(t)

31
which is a linear diffusion.

18 Forward Rate Models

Exercise 18.1 We know that the Hull-White model

dr(t) = (Θ(t) − ar(t))dt + σdW (t)

has an affine term structure, i.e., that

p(t, T ) = eA(t,T )−B(t,T )r(t) ,

1


and that in this model we have in particular, B(t, T ) = a 1 − e−a(T −t) .
∂ ln p(t,T )
Furthermore, f (t, T ) = − ∂T = − ∂A(t,T
∂T
)
+ ∂B(t,T )
∂T r(t), so in this case we
have
∂A(t, T )
f (t, T ) = − + e−a(T −t) r(t).
∂T
By Itô we have that the dynamics under Q of the forward rates is given by
 
∂ ∂A(t, T )
df (t, T ) = − dt + ae−a(T −t) r(t)dt + e−a(T −t) dr(t)
∂t ∂T
 
∂ ∂A(t, T )
= − dt + ae−a(T −t) r(t)dt + e−a(T −t) [(Θ(t) − ar(t))dt + σdW (t)]
∂t ∂T
= α(t, T )dt + e−a(T −t) σdW (t),
h i
∂ ∂A(t,T )
where α(t, T ) = − ∂t ∂T + e−a(T −t) Θ(t).

Exercise 18.2 Take as given an HJM model (under Q) of the form

df (t, T ) = α(t, T )dt + σ(t, T )dW (t),

where the volatility σ(t, T ) is a deterministic function of t and T .

RT
(a) By the HJM drift condition, α(t, T ) = σ(t, T ) t σ 0 (t, u)du ∀t, T , which
for deterministic σ(t, T ), means that α(t, T ) is also deterministic.
To see the distribution of the forward rates note that
Z t Z t
f (t, T ) = f (0, T ) + α(s, T )ds + σ(s, T )dW (s)
0 0
Z t Z T Z t
0
= f (0, T ) + σ(s, T ) σ (s, u)duds + σ(s, T )dW (s).
0 s 0
| {z }
µ(t,T )

32
 Note that f (0, T ) is observable and that the double integral is determin-
Rt
istic, so the only stochastic part is 0 σ(s, T )dW (s).
h Rt i
Hence f (t, T ) ∼ N µ(t, T ), 0 σ 2 (s, T )ds .
Since r(t) = f (t, t) we immediately have that r(t) is also normally distrib-
uted.
n R o
T
(b) Since we have p(t, T ) = exp − t f (t, s)ds , and in (a) we have already
shown that f (t, s) is normally distributed, then p(t, T ) is lognormally dis-
tributed.

Exercise 18.3 Recall that in between the vectors of market-price of risks of

the domestic market and the foreign market there is the following relation:

λf (t) = λd (t) − σx0 (t).

where, σx (t) is the vector of volatilities of in the SDE for the exchange-rate
X(denoted in units of domestic currency per unit of foreign currency).

We also know that any process, so in particular forward rates, has under Q (the
domestic martingale measure), the drift term equal to (µ(t, T ) − σ(t, T )λd (t)),
where µ(t, T ) is the drift term under the objective probability measure and that
the volatility term remains the same. So, in particular, for the foreign forward
rates must have under the domestic martingale measure, Q,

dff (t, T ) = (µf (t, T ) − σf (t, T )λd (t)) dt + σf (t, T )dW (t),

hence, αf (t, T ) = µf (t, T ) − σf (t, T )λd (t).

Likewise, any process, under Qf (the foreign martingale measure),has a drift

term equal to (µ(t, T ) − σ(t, T )λf (t)).

So, in particular, for the foreign forward rates we have under Qf

dff (t, T ) = (µf (t, T ) − σf (t, T )λf (t)) dt + σf (t, T )dW f (t).
| {z }
α̃f (t,T )

So, using the relation between λd and λf we can also establish a relation between
αf and α̃f ,

α̃f (t, T ) = µf (t, T ) − σf (t, T )λf (t)

= µf (t, T ) − σf (t, T ) [λd (t) − σx0 (t)]
= µf (t, T ) − σf (t, T )λd (t) +σf (t, T )σx0 (t).
| {z }
αf (t,T )

33
We also have that, under the foreign martingale measure, Qf , the coefficients of
the foreign martingale measure must satisfy the standard HJM drift condition
so: Z T
α̃f (t, T ) = σf (t, T ) σf0 (t, s)ds.
t
Using teh relation found between αf and α̃f and using the drift condition above
we get
Z T
α̃f (t, T ) = σf (t, T ) σf0 (t, s)ds
t
Z T
αf (t, T ) + σf (t, T )σx0 (t) = σf (t, T ) σf0 (t, s)ds
"tZ #
T
αf (t, T ) = σf (t, T ) σf0 (t, s)ds − σx0 (t) .
t

19 Change of Numeraire

Exercise 19.1 In the Ho-Lee model we have under Q

dr(t) = Θ(t)r(t)dt + σdW (t)

And we know that on this model we have an affine term structure for bond-prices

p(t, T ) = eA(t,T )−B(t,T )r(t)

with B(t, T ) = T − t.

If the Q-dynamics of p(t, T ) are given by

dp(t, T ) = r(t)p(t, T )dt + v(t, T )p(t, T )dW (t)

then by Ito we have

v(t, T ) = −σB(t, T ) = −σ(T − t).

The project is to price an European call option with:

• date of maturity T1

• strike price K

• where the underlying is a zero-coupon bond with date of maturity T2

34
and T1 < T2 .

Note that our Z-claim is given by

Z = max [p(T1 , T2 ) − K; 0] ,

Hence, we have using a change of measure on the standard arbitrage pricing

formula that
 RT 
1
− r(s)ds
Π(t; Z) = EtQ e t max [p(T1 , T2 ) − K; 0]

= p(t, T1 )EtT1 [max [p(T1 , T2 ) − K; 0]]

  
  p(T , T ) 
  1 2 
= p(t, T1 )EtT1 max  −K; 0 .
  p(T1 , T1 ) 
| {z }
Z(T1 )

Note that dividing p(T1 , T1 ) in the last step of the equation “changes noth-
ing”(since p(T, T ) = 1 for all T ) but has the advantage of seeing our claim as
a claim on the process Z, which we know is a martingale under QT1 , and as
long as it has deterministic volatility the Black-Scholes formula can help us. To
check this note that
p(t, T2 )
Z(t) =
p(t, T1 )
can also be written as (just using the fact that we have an ATS)

Z(t) = exp {A(t, T2 ) − A(t, T1 ) − [B(t, T2 ) − B(t, T1 )] r(t)} ,

and, therefore, has the following dynamics under Q(applying Ito formula)

dZ(t) = {· · ·} Z(t)dt + Z(t)σz (t)dW (t)

where σz (t) = −σ [B(t, T2 ) − B(t, T1 )] = −σ (T2 − T1 ), will be the same as

under QT1 , and is deterministic.

Since under QT1 ,

Z T1
Z(T1 ) = Z(t) − σ (T2 − T1 ) dW (s)
t
h i
2
The conditional (on information at time t) distribution of Z is Z(T1 ) ∼ N Z(t), σ 2 (T2 − T1 ) (T1 − t) .

So, (from the BS formula)

Π(t) = p(t, T1 ) {Z(t)N [d1 (t, T1 )] − KN [d2 (t, T1 ]}

= p(t, T2 )N [d1 (t, T1 )] − p(t, T1 )KN [d2 (t, T1 ]

35
where
ln Z(t)
2
+ 12 σ 2 (T2 − T1 ) (T1 − t)
p(t,T2 )
ln Kp(t,T 1)
+ 12 σ 2 (T2 − T1 )2 (T1 − t)
K
q q
d1 (t, T1 ) = =
2 2
σ 2 (T2 − T1 ) (T1 − t) σ 2 (T2 − T1 ) (T1 − t)
q
2
and d2 (t, T1 ) = d1 (t, T1 ) − σ 2 (T2 − T1 ) (T1 − t).

Exercise 19.2 Take as given an HJM model of the form

df (t, T ) = α(t, T ) + σ(t, T )dW (t) + σ2 e−a(T −t) dW2 (t)

 
W1 (t)
where σ(t, T ) = [ σ1 (T − t) σ2 e−a(T −t) ], W (t) = and we have W1
W2 (t)
and W2 independent Wiener processes and σ1 and σ2 constants.

(a) From the relation between the dynamics of bond prices and forward rates
we know that

dp(t, T ) = r(t)p(t, T )dt + p(t, T )S(t, T )dW (t)

RT
with S(t, T ) = − t σ(t, s)ds.
In this case we have then
Z T
S(t, T ) = − [ σ1 (s − t) σ2 e−a(s−t) ] ds
σ2  #
t
" σ1
− (T − t)2 − 1 − e−a(T −t)
= | 2 {z } | a {z }
σ1p (t,T ) σ2p (t,T )

So,

dp(t, T ) = r(t)p(t, T )dt + p(t, T )S(t, T )dW (t)

 

dW1 (t)
= r(t)p(t, T )dt + p(t, T ) [ − σ21 (T − t) 2
− σa2 1−e −a(T −t)
]
dW2 (t)
 
dW1 (t)
= r(t)p(t, T )dt + p(t, T ) [ σ1p (t, T ) σ2p (t, T ) ]
dW2 (t)

(b) We will use exactly the same technique as in exercise 19.1 to price an
European call option with maturity T0 on a T1 -bond.
p(t,T1 )
Note that in this model the Z process Z(t) = p(t,T0 ) is a martingale under
T0
Q and its dynamics has deterministic volatility given by
Z T1
σT1 ,T0 (t) = S(t, T1 ) − S(t, T0 ) = − [ σ1 (s − t) σ2 e−a(s−t) ] ds
T0

36
Hence (by the same arguments as in exercise 19.1)
 
 Z T0 
 

Z(T0 ) ∼ N Z(t), 2 
||σT1 ,T0 (s)|| ds
 |t {z }
Σ2T
0 ,T1

and the pricing formula is given by

Π(t) = p(t, T0 ) {Z(t)N [d1 (t, T0 )] − KN [d2 (t, T0 ]}

= {p(t, T1 )N [d1 (t, T0 )] − p(t, T0 )KN [d2 (t, T0 ]}

where
p(t,T2 )
ln Z(t) 1 2
K + 2 ΣT0 ,T1
ln Kp(t,T 1)
+ 12 Σ2T0 ,T1
d1 (t, T0 ) = q = q
Σ2T0 ,T1 Σ2T0 ,T1
q
and d2 (t, T1 ) = d1 (t, T1 ) − Σ2T0 ,T1 .

37
20 Extra Exercises

20.1 Exercises

Exercise 20.1 Exercise 3.4 in the book!

Exercise 20.2 Solve explictly the GMB SDE

dX(t) = µX(t)ds + σXdW (t),

X(t) = x.

Exercise 20.3 Let {Zn } be a sequence of i.i.d. (independent identically dis-

tributed) random variables with finite exponential moments of all orders. Define
the function ϕ : R → R by
 
ϕ(λ) = E eλZn ,

and define the process X by

n
X
eλSn
Xn = n, where Sn = Zk .
[ϕ(λ)]
k=1

Prove that X is an Fn -martingale, where Fn = σ {Zi ; i = 1, . . . , n}.

Exercise 20.4 Compute the infitesimal generator of the following known processes

(a) Brownian Motion

dY (t) = dW (t)

(b) Geometric Brownian Motion

dS(t) = µS(t)dt + σS(t)dW (t)

(c) Vasicek Model for the short rate

dr(t) = c (µ − r(t)) dt + σdW (t)

(d) The Ornstein-Uhlenbeck process

dX(t) = −αX(t)dt + σdW (t)

38
 (e) Graph of the Brownian Motion

dX1 (t) = dt X1 (0) = t0
dX2 (t) = dB(t) X2 (0) = 

Exercise 20.5 Exercise 6.3 in the book.

Derive the Black-Scholes formula for the standard call option.

Exercise 20.6 Derive the formulas of the greeks of an European call with strike
price K and time of maturity T . That is, prove that following relations, (with
notation as in the Black–Scholes formula and where the letter ϕ denotes the
density function of the N [0, 1] distribution,) holds

∆ = N (d1 ),
ϕ(d1 )
Γ = √ ,
sσ T − t
ρ = K(T − t)e−r(T −t) N (d2 ),
sϕ(d1 )σ
Θ = − √ − rKe−r(T −t) N (d2 ),
2 T −t
√
V = sϕ(d1 ) T − t.

Hint: Prove that sϕ(d1 ) = Ke−r(T −t) ϕ(d2 ), and use this to prove the results.

Exercise 20.7 Exercise 11.5 in the book.

Exercise 20.8 Derive the formula for the value of a call option on a stock in a
world were we do not assume deterministic interest rates.

20.2 Solution

Exercise 20.1 Integrating we get

Z t Z t
X(t) = X(0) + α X(s)ds + σ(s)dW (s)
0 0

Taking expectations we have

Z t
E [X(t)] = E [X(0)] + α E [X(s)] ds
0

By setting m(t) = E [X(t)] and after some calculations we obtain the answer

E [X(t)] = eαt E [X(0)]

39
Exercise 20.2 Take Y (t) = ln X(t), then by Itô we have
 
1
dY (t) = µ − σ 2 dt + σdW (t)
2
Integrating we get for T > t:
Z T  Z T
1
Y (T ) = Y (t) + µ − σ 2 ds + σdW (t)
t 2 t
 
1
= Y (t) + µ − σ 2 (T − t) + σ(W (T ) − W (t))
2
In terms of X we have
 
1
ln X(t) + µ − σ 2 (T − t) + σ(W (T ) − W (t))
ln X(T ) =
2
  
1
X(T ) = |{z}
x exp µ − σ 2 (T − t) + σ(W (T ) − W (t)) .
2
X(t)

Exercise 20.3 We have

n
X
  eλSn
ϕ(λ) = E eλZn , Xn = n, Sn = Zk .
[ϕ(λ)]
k=1

We want to show that

E [ Xn+1 | Fn ] = Xn ⇔
|{z} E [ Xn+1 − Xn | Fn ] = 0.
Since Xn ∈ Fn

From X definition we have

eλSn+1 eλSn
Xn+1 − Xn = −
[ϕ(λ)]
n+1 [ϕ(λ)]n
Pn+1 Pn
eλ k=1 Zk − ϕ(λ)eλ k=1 Zk
= n
[ϕ(λ)]
Pn
eλ k=1 Zk λZn+1

= n e − ϕ(λ) .
[ϕ(λ)]
| {z }
Xn

Hence

 
E [ Xn+1 − Xn | Fn ] = E Xn eλZn+1 − ϕ(λ)  Fn
   
= Xn E eλZn+1 − E eλZn  Fn
     

= Xn E eλZn+1  Fn − E eλZn  Fn
| {z }
0, since {Zn } are iid
= 0 and we conclude that X is an Fn martingale.

40
Exercise 20.4

2
1∂ f
(a) Af (x) = 2 ∂x2 (x).
2
(b) Af (x) = µx ∂f 1 2 2∂ f
∂x + 2 σ x ∂x2 (x).
2
(c) Af (x) = 12 σ 2 ∂∂xf2 (x) + c (µ − x) ∂f
∂x .
2
(d) Af (x) = −αx ∂f 1 2∂ f
∂x + 2 σ ∂x2 (x).
 
1
(e) In vector notation we have dX(t) = bdt + σdW (t), where b = and
  0
0 0
σ= . And it follows that
0 1

∂f 1 ∂2f
Af (x1 , x2 ) = (x1 , x2 ) + (x1 , x2 ),
∂x1 2 ∂x2 2
or since X1 = t

∂f 1 ∂2f
Af (t, x) = (t, x) + (t, x).
∂t 2 ∂x2

Exercise 20.5 In the case of a standard call option on a stock with price
process S with maturity T and exercise price K, we have that the option’s
price at time t and given S(t) = s, is

Q
c(t, s) = e−r(T −t) Et,s [max (S(T ) − K, 0)] . (1)

where S has the following Q-dynamics

dS(t) = rS(t)dt + σS(t)dW (t).

In order to compute the expectation in equation (1), recal that since S follows
a GBM we get
 
1 2
S(T ) = S(t) exp (r − σ )(T − t) + σ(W (T ) − W (t)) .
|{z} 2
s

˜ )
So, S(T ) has the same distributions of the the normalized variable S(T
 
1 √
S̃(T ) = s exp (r − σ 2 )(T − t) + σ T − t Z ,
2

where Z ∼ N (0, 1).

41
 h  i
Q Q
And hence we have Et,s [max (S(T ) − K, 0)] = Et,s max S̃(T ) − K, 0 .

Note also
h  i h i h i
Q Q Q
Et,s max S̃(T ) − K, 0 = Et,s S̃(T )|S̃(T ) ≥ K −K Et,s 1|S̃(T ) ≥ K .
| {z } | {z }
A B

Starting with part B we see that

h i h i
Q
B = Et,s 1|S̃(T ) ≥ K = K Q S̃(T ) ≥ K
 


 ln Ks − (r − 12 σ 2 )(T − t) 

= K Q Z ≥ √  = K N [−Z0 ]
 σ T −t 
| {z }
Z0

where N is the cumulative distribution function (CDF) of the standard normal

distribution with density
Z x
1 2
φ(z) = √ e−z /2 . So, we have N [x] = φ(z)dz.
2Π −∞

Taking now part A, we have

h i Z ∞ √
Q 1 2
Et,s S̃(T )|S̃(T ) ≥ K = se(r− 2 σ )(T −t)+σ T −t z φ(z)dz
Z0
Z ∞ √
1 2 1 2
= se(r− 2 σ )(T −t)+σ T −t z √ e−z /2 dz
Z0 2Π
Z ∞ √
1 1 2 2
= ser(T −t) √ e− 2 σ (T −t)+σ T −t z−z /2 dz
Z0 2Π
Z ∞ √
1 (z−σ T −t)2
r(T −t)
= se √ e− 2 dz
Z0 2Π
| {z }
√
density of a N (σ T − t, 1)
= {which using the CDF of a standard normal, N [x], becomes}
 h √ i
= ser(T −t) 1 − N Z0 − σ T − t
 h √ i
= ser(T −t) N −Z0 + σ T − t
= ser(T −t) (N [d1 (t, T )])

where
K


√ ln − (r − 12 σ 2 )(T − t)
s
√
d1 (t, T ) = −Z0 + σ T − t = − √ +σ T −t
σ T −t


ln K + (r + 12 σ 2 )(T − t)
s
= √
σ T −t

42
And if also define
s


√ ln K + (r − 12 σ 2 )(T − t)
d2 (t, T ) = −Z0 = d1 (t, T ) − σ T − t = √ ,
σ T −t
we finally get the Black-Scholes formula (with d1 and d2 as just defined)
n o
c(t, s) = e−r(T −t) ser(T −t) N [d1 (t, T )] − K N [d2 (t, T )]
= sN [d1 (t, T )] − K e−r(T −t) N [d2 (t, T )] .

Exercise 20.6 Take the Black-Scholes formula:

c(t, s) = sN [d1 (t, s)] − Ke−r(T −t) N [d2 (t, s)]

where    
1 S 1
d1 (t, s) = √
ln + r + σ 2 (T − t)
σ T −t K 2
√
and d2 (t, s) = d1 (t, s) − σ T − t

Using the hint let us first prove that

sϕ(d1 ) = Ke−r(T −t) ϕ(d2 ) (2)

holds where ϕ is the density function of the N [0, 1] distribution. So we have

d2
1
ϕ(d1 ) = √1 e− 2 and
2π

d2
1 2
ϕ(d2 ) = √ e− 2
2π
√ 2
1 − (d1 −σ T −t)
= √ e 2
2π
1 − d21 − σ2 (T −t) −d1 σ√T −t
= √ e 2 e 2 e
2π
| {z }
ϕ(d1 )

So it remains to show that

σ2 (T −t) √ K −r(T −t)
e− 2 e−d1 σ T −t
= e
s
taking logs and rearranging we get

σ 2 (T − t) √ K
− − d1 σ T − t = ln (−r(T − t))
2 s
σ 2 (T −t)
ln K
s + (r(T − t)) − 2
d1 = √
σ T −t

43
which is true by definition of d1 and concludes the proof.

With the relation (2) in mind we can now go for the task of calculating the
greeks of a call option.

• Delta
The delta measures the impact in the value of the derivative that results
from price variations of the underlying, differentiating w.r.t s the Black-
Scholes formula, we have
∂c(t, s) ∂N [d1 ] ∂d1 (t, s) ∂N [d2 ] ∂d2 (t, s)
∆c = = N [d1 ] + s − Ke−r(T −t) .
∂s ∂d1 ∂s ∂d2 ∂s
∂N [d1 ] ∂N [d2 ]
Note that ∂d1 = ϕ(d1 ) and likewise ∂d2 = ϕ(d2 ) and we also have
that
∂d1 (t, s) ∂d2 (t, s) 1
= = √
∂s ∂s sσ T − t
Joining the pieces and using equation (2) we get the result:
1 1
∆c √
= N [d1 ] + sϕ(d1 ) − Ke−r(T −t) ϕ(d2 ) √
sσ T − t sσ T − t
1 h i
= N [d1 ] + √ sϕ(d1 ) − Ke−r(T −t) ϕ(d2 )
sσ T − t | {z }
0

• Gamma
The gamma measures the impact in the value of the delta that results
from price variations of the underlying, differentiating w.r.t s ∆c found
above we get, immediately get the result:
∂∆c ∂N [d1 ] ∂d1 (t, s) 1
= = ϕ(d1 ) √ .
∂s ∂d1 ∂s sσ T − t

• Rho
The rho measures the impact in the value of the derivative that results
from interest rate variations, differentiating w.r.t r, the Black-Scholes for-
mula we get
∂d1 ∂d2
ρc = Sϕ(d1 ) + K(T − t)e−r(T −t) N [d2 ] − Ke−r(T −t) ϕ(d2 )
∂r ∂r
Since ∂d ∂d2
∂r = ∂r and from the relation (2) the first and third parcels on
1

the RHS cancel the result follows

ρc = K(T − t)e−r(T −t) N [d2 ] .

44
 • Theta
The theta measures the impact in the value of the derivative that results
from the decrease in time to maturity, differentiation w.r.t ”time to ma-
turity” is the same as differentiation w.r.t. −T . So, differentiating the
Black-Scholes formula w.r.t. (-T) we get
 
∂d1 ∂d2
Θc = − Sϕ(d1 ) + Ke−r(T −t) N [d2 ] − Ke−r(T −t) ϕ(d2 )
∂T ∂T
∂d2 ∂d1 √1
Since ∂T = ∂T − 2 T −t
and using the relation (2) we get
  
∂d1 ∂d1 1
Θc = − Sϕ(d1 ) + Ke−r(T −t) N [d2 ] − Ke−r(T −t)ϕ(d2 ) − √
∂T ∂T 2 T −t
Sϕ(d1
= − √ − rKe−r(T −t) N [d2 ] .
2 T −t

• Vega
The vega measures the impact in the value of the derivative that results
from volatility variations, differentiating w.r.t σ, the Black-Scholes formula
we get
∂d1 ∂d2
νc = Sϕ(d1 ) − Ke−r(T −t)ϕ(d2 )
∂r ∂r
∂d2 ∂d1
√
Since ∂r = ∂r − T − t and from the relation (2) result follows
√
νc = Sϕ(d1 ) T − t.

Exercise 20.7 We have to show that Fδ (t, s) = F0 (t, se−δ(T −t) ), i.e, that the
pricing function when we have a continuous dividend process δ is just the “stan-
dard” (no-dividend) price function where we replace s by se−δ(T −t) . To see this,
let us denote S the price of a stock that pays a continuous dividend δ, and S̄
the price of an otherwise equivalent stock but that does not pay any dividend.

Q  
F0 (t, s) = e−r(T −t) Et,s Φ(S̄(T ))
= {with dS̄(t) = rS̄(t)dt + σS̄(t)dW Q (t)}
  
S̄(T )
 z }| { 
Q   (r− 12 σ 2 )(T −t)+σ(W (T )−W (t) 
= e−r(T −t) Et,s Φ
  se  (3)

45
 Q
Fδ (t, s) = e−r(T −t) Et,s [Φ(S(T ))]
= {with dS(t) = (r − δ) S(t)dt + σS(t)dW Q (t)}
  
S(T )
  z }| { 
Q   (r−δ− 12 σ 2 )(T −t)+σ(W (T )−W (t)) 
= e−r(T −t) Et,s Φ se  (4)

By comparing the arguments of the function Φ(·) in (3) and (4) we immediately
get the result since,
h  1 2
i
Q
F0 (t, se−δ(T −t) ) = e−r(T −t) Et,s Φ e−δ(T −t) se(r− 2 σ )(T −t)+σ(W (T )−W (t))
  
S̄(T )
  z }| {
Q   (r−δ− 12 σ 2 )(T −t)+σ(W (T )−W (t) 
= e−r(T −t) Et,s Φ se 

= Fδ (t, s)

Exercise 20.8

X = max [S(T ) − K; 0]
= [S(T ) − K] .I {S(T ) ≥ K}

where 
1 if S(T ) ≥ T
I {S(T ) ≥ K} =
0 if S(T ) < K
We obtain,
 RT 
− r(s)ds
Π(0, X ) = EtQ
e 0 [S(T ) − K] .I {S(T ) ≥ K}
 RT   RT 
Q − r(s)ds Q − r(s)ds
= Et e 0 S(T ).I {S(T ) ≥ K} − Et e 0 K.I {S(T ) ≥ K}

= {we change the measure on the first term to QS , where S is the numeraire}

{for the second term we change measure but take the T -zero-coupon bond as the numeraire}

= S(0)Q (S(T ) ≥ K) − Kp(0, T )QT (S(T ) ≥ K)

S

Take first the second term. We have

 
 S(T ) 
 
QT (S(T ) ≥ K) = QT  ≥ K  = QT (ZS,T (T ) ≥ K)
 p(T, T ) 
| {z }
ZS,T (T )

46
But, under the T -forward measure we have

dZS,T (t) = ZS,T (t)σS,T (t)dW T

This is a basic GBM hence,

( Z Z )
T T
1 2 T
ZS,T (T ) = ZS,T (0) exp − kσS,T k (t)dt + σS,T (t)dW .
2 0 0

Which means that

QT (ZS,T ≥ K) = QT (ln(ZS,T (T )) ≥ ln(K))

   

 

   RT 
1 2
= ln(ZS,T (T ))|Ft ∼ N ln(ZS,T (0)) − Σ , Σ2  Σ2S,T = kσS,T k2 (t)dt
  2 S,T S,T  0 

 | {z } 

µ
 
ln K − µ 
= 1−N q
Σ2S,T )
 
 
 
 − ln K + µ 
= N
 q 

 Σ2S,T 
 
| {z }
d2

For the first term, we have

 
   p(T, T )  
1 1  1
 1
QS (S(T ) ≥ K) = QS ≤ = QS  ≤  = QS YS,T (T ) ≤
S(T ) K  S(T ) K K
| {z }
YS,T (T )

1
That is YS,T (t) = ZS,T (t). Since the process YS,T is a normalized variable when
S is the numeraire we have, under QS

dYS,T (t) = YS,T (t)δS,T dW S

and from it is easy to show (why?) that δS,T = −σS,T , thus we have
( Z Z T )
1 T 2 S
YS,T (T ) = YS,T (0) exp − kσS,T k (t)dt − σS,T (t)dW .
2 0 0

47
So,
   
1 1
QS YS,T (T ) ≤ = QS ln(YS,T (T )) ≤ ln( )
K K
   

 

 1 2  RT
= ln(YS,T (T ))|Ft ∼ N ln(YS,T (0)) − Σ , Σ2  Σ2S,T = kσS,T k2 (t)dt

 2 S,T S,T 0 

| {z }
α
 
− ln K − α 
= N q
Σ2S,T
= {note that α = −µ − ΣS,T }
 
 
 
 − ln K + µ + Σ2S,T 

= N q 

 Σ2S,T 
 
| {z }
d1

q
and we see that actually we have d1 = d2 + Σ2S,T .

The price of a call option in a stochastic interest rate world is thus,

Π(0) = S(0)N [d1 ] Kp(0, T )N [d2 ]

with d1 and d2 as defined above!

48
21 Exams

21.1 Exercises

21.1.1 March 26, 2003

Exercise 21.1 Let Y be given as the solution to the following SDE.

dX = αX(t)dt + σX γ dW

here α, γ and σ are deterministic constants.

(a) Compute E [X(t)].

(b) Compute V ar[X(t)] for the case when γ = 1.

 
Hint: Start by computing E X 2 (t)

Exercise 21.2 The well known company Cathy and Heathcliff INC. are trad-
ing the new derivative, “the Fraction”, which is defined by two fixed dates T0
and T1 with T0 < T1 . The holder of a Fraction contract will, at the date T1
obtain the amount
S(T1 )
X=
S(T0 )

(a) Determine the arbitrage free price, at time t < T0 of the Fraction contract.
You live in a standard Black-Scholes world with short rate r.

(b) Try to construct the replicating portfolio for the Fraction contract. (It can
be done almost by inspection).

Exercise 21.3 Consider the following SDE under the objective probability
measure P .
dXt = µ(Xt )dt + σ(Xt )dWt

where µ(x) and σ(x) are deterministic given functions. We interpret X as a

non priced underlying object, and now we want to price derivatives defined in
terms of X. One possibility is to interpret Xt as the temperature (at time t) at
a particular beach at Tylösand. Now consider the contingent claim

Z = Φ(XT )

49
where Φ(x) is some given deterministic function. In our interpetation this could
be interpreted as a holidy insurance, which gives the holder an amount of money
if the temperature is below some benchmark value.

(a) Is it possible to obtain a unique arbitrage free price process Π (t; Z) for the
derivative above? You must motivate your answer.

(b) Use the standard ideas from interest rate theory to analyze the situation
above as completely as possible. In particular you should come up with a
PDE for pricing of derivatives. Indicate clearly which objects in the PDE
that are known to you, and how it in priciple would be possible to obtain
further necessary information.

Exercise 21.4 Consider a standard Black-Scholes model under the objective

probability measure P :

dS(t) = αS(t)dt + σS(t)dWt

dB(t) = rB(t)dt.

Consider two fixed points in time, T1 and T2 with T1 < T2 and define the
T2 -claim X by
X = max [ST1 , ST2 ] .
Compute the arbitrage free price Π (t; X) of X at time t where t ≤ T1 . You are
allowed to use the standard Black-Scholes formula without proof (see below).

Exercise 21.5 Consider a standard Black-Scholes model. Show how it is pos-

sible to hedge (replicate) a given contingent claim X of the form X = Φ(S(T )).

Black-Scholes formula:

F (t, s) = sN [d1 (t, s)] − e−r(T −t) KN [d2 (t, s)] .

Here N is the cumulative distribution function for the N [0, 1] distribution and
     
1 s 1
d1 (t, s) = √ ln + r + σ 2 (T − t) ,
σ T −t K 2
√
d2 (t, s) = d1 (t, s) − σ T − t.

50
21.1.2 January 10, 10.00-15.00, 2003

All notation should be clearly defined.

Arguments should be complete and careful.
Each problem below will give you a maximum of 20 points.

Exercise 21.6 Let X and Y be given as solutions of the following system of

SDEs

dX = αXdt + σXdW, X 0 = x0
dY = βY dt + γY dW, Y0 = y0

Note that both equations are driven by the same scalar W .

(a) Define Z by Zt = Xt Yt . It turns out that Z satisfies an SDE itself. Which?

(b) Compute E[Zt ].

Exercise 21.7 The rather unknown company William S. INC has blessed the
market with the new derivative, ”As You Like It” (AYLI). The AYLI is defined
in terms of an underlying stock with price process S, and it is specified by two
dates, T0 and T1 with T0 < T1 and an exercise price K. The holder of the AYLI
may at time T0 choose th have either a call option on S with strike price K and
exercise date T1 or a put option on S with strike price K and exercise date T1 .
Derive a formula, for the price at time t < T0 for the AYLI. The stock price is
assumed to follow standard GBM

dS = αSdt + σSdW,

and the short rate r is assumed to be constant. If, for some reason, you want to
use the Black-Scholes formula for a call option, you may do so without proof.

Exercise 21.8 Consider a bond market with bond prices as usual denoted by
p(t, T ).

(a) Define the forward rates f (t, T ) in terms of bond prices, and derive a for-
mula expressing bond prices in terms of forward rates.

51
(b) Assume that the forward rates under Q have the dynamics

df (t, T ) = α(t, T )dt + σ(t, T )dWt

and derive the relation which must hold between α and σ.

Exercise 21.9 Specify the assumptions in the Black-Scholes model, and derive
the Black-Scholes partial differential equation for the price of a simple claim of
the form X = Φ(ST ).

Exercise 21.10 Consider the following model for two stock prices (without
dividends)

dS1 = α1 S1 dt + S1 σ1 dW,
dS2 = α2 S2 dt + S2 σ2 dW.

Here α1 and α2 are known real numbers, whereas σ1 and σ2 are two-dimensional
constant row vectors. W is a standard two dimensional Wiener process. The
short rate is assumed to be constant. The models is assumed to be arbitrage
free. Consider the T -claim X defined by the following points:

• X is specified in terms of the underlying stocks , a strike price K, and an

exercise date T .

• The holder of the claim will obtain nothing if S2 (T ) ≤ K.

• If S2 (T ) > K then the holder of the claim will obtain S2 (T ) − K shares

(i.e. units) of asset Number 1.

Compute the price Π (t; X).

52
21.1.3 March 18, 2002

All notation should be clearly defined. Arguments should be complete and care-
ful.

Exercise 21.11 Consider the following boundary value problem in the domain
[0, T ] × R for an unknown function F (t, x).

∂F ∂F 1 ∂2F
+ µ(t, x) + σ 2 (t, x) 2 + k(t, x)F (t, x) = 0,
∂t ∂x 2 ∂x
F (T, x) = Φ(x).

Here µ, σ, k and Φ are assumed to be known functions.

Derive a Feynman-Kač representation formula for this problem. In this formula

it must be quite clear exactly at which points the various functions should be
evaluated. In other words - if you suppress variables you must explain very
clearly exactly what you mean.

If you think this problem is hard you may (with loss of 10 points) assume that
k(t, x) = 0. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (40p)

Exercise 21.12 Assume that the short rate under the objective measure P has
the dynamics
dr = µ(t, r)dt + σ(t, r)dW (t)

Derive the “term structure equation” for the determination of bond prices. (40p)

Exercise 21.13 Consider a standard Black-Scholes model under the objective

probability measure P :

dS(t) = αS(t)dt + σS(t)dWt

dB(t) = rB(t)dt.

We want to price a European call option with exercise price K and exercise date
T on the underlying stock. Let us also assume that the stock pays a continuous
dividend yield δ, i.e. that the cumulative dividend stream has the structure

dDt = δSt dt

where δ > 0.

53
(a) Discuss verbally what relation you would expect to hold between the option
price in the case of a dividend yield compared to the option price in the
case when the dividend yield is zero. (Which would be greater and why?)
(5p)

(b) Derive a pricing formula (as explicit as possible) for the call option in the
case of a nonzero dividend yield. You are allowed to use and refer to the
standard Black-Scholes formula without proof (see below). . . . . . . . . (35p)

Exercise 21.14 Consider two markets: a domestic market with short rate rd ,
and a foreign market with short rate rf . The exchange rate X is defined as the
domestic price of one unit of the foreign currency.

We take as given the following dynamics (under the objective probability mea-
sure P )

dX = XαX dt + XσX dW, (5)

dBd = rd Bd dt, (6)
dBf = rf Bf dt, (7)

where rd , rf ,αX , σX are deterministic constants, and W is a scalar Wiener

process.

We want to price (in domestic terms) a European Call Option on one unit of
the foreign currency, with strike price K and exercise date T . The contract Y
is thus given by
Y = max[X(T ) − K, 0]
Derive an explicit pricing formula for the option. The standard Black-Scholes
formula for stock options (see below) can be used without further motivation.
(40p)

Exercise 21.15 Consider a standard Black-Scholes model under the objective

probability measure P :

dS(t) = αS(t)dt + σS(t)dWt

dB(t) = rB(t)dt.

Consider two fixed points in time, T1 and T2 with T1 < T2 and define the
T2 -claim X by
X = max [ST1 , ST2 ] .

54
Compute the arbitrage free price Π (t; X) of X at time t where t ≤ T1 . You are
allowed to use the standard Black-Scholes formula without proof (see below).
(40p)

Black-Scholes formula:

F (t, s) = sN [d1 (t, s)] − e−r(T −t) KN [d2 (t, s)] .

Here N is the cumulative distribution function for the N [0, 1] distribution and
     
1 s 1 2
d1 (t, s) = √ ln + r + σ (T − t) ,
σ T −t K 2
√
d2 (t, s) = d1 (t, s) − σ T − t.

21.2 Topics of solutions

21.2.1 March 26, 2003

Exercise 21.1

(a) Using standard technique and using the notation mt = E [Xt ] we obtain

dmt
= αmt ,
dt
m0 = X0

and we get
E [Xt ] = mt = X0 eαt .

(b) There are several ways to solve this, but an easy one is to compute dX 2
as


dX 2 = X 2 2α + σ 2 dt + 2σX 2 dW

This gives us
  2
E Xt2 = X02 e(2α+σ )t

so  2 
2
2
V ar[Xt ] = X02 e(2α+σ )t
− X0 eαt = X02 e2αt eσ t − 1

Exercise 21.2

55
 (a) Since we have a standard Black-Scholes model it is easily seen that
1 2
X = e(r− 2 σ )(T1 −T0 )+σ[W (T1 )−W (T0 )]
,

The price is given as usual by

Q
Π (t; X) = e−r(T1 −t) Et,s [X]

and from the formula for X above we immediately have

Q
Et,s [X] = er(T1 −T0 )

so the answer is
Π (t; X) = e−r(T0 −t) .

(b) From (a) we know that Π (0; X) = e−rT0 , and the replicating portfolio is
as follows.

• At t = 0 put e−rT0 SEK into the bank.

• At t = T0 we have 1 SEK in the bank. For this amount we can buy
1/S(T0 ) shares.
• At t = T1 each share is worth S(T1 ) SEK so the value of the portfolio
is S(T1 )/S(T0 ) SEK.

Exercise 21.3 See the textbook, Chapter 10.

Exercise 21.4 The hard part about this problem is that ST1 is not the price of
a traded asset at the time T2 . Therefore we cannot immediately apply the change
of numeraire technique. In order to transform our problem to the standard
setting we therefore introduce a self financing portfolio with value process V .
This portfolio is defined as follows, for a fixed t < T1 :

• Start at t with one share of the stock.

• Hold this portfolio until T1 . At T1 sell it, and put all the money in the
bank.

• Keep the money in the bank until T2 .

56
The point is that V can be regarded as the price of a traded asset. An easy
calculation shows that VT2 = er∆ST1 where ∆ = T2 − T1 . Thus we have ST1 =
e−r∆ VT2 and we can write
 
X = max e−r∆ VT2 , ST2

Using V as the numeraire we now obtain (see Ch. 19 for details)

V
  
Π (t; X) = Vt · Et,s max ZT2 , e−r∆

Su
where Zu = Vu and E V denotes the martingale measure with V as the nu-
meraire. An easy calculation gives us
  
Π (t; X) = Vt e−r∆ + Vt · Et,s
V
max ZT2 − e−r∆, 0

and in the last expectation we recognize a European call on Z with strike e−r∆ .
The normailzed price system has (as usual) zero interest rate, so this option can
be valued using the BS formula and we only have to compute the volatility σZ
of Z. This volatility is time dependent, and from the definition of V it follows
immediately that (
0 for t ≤ u ≤ T1
σZ (u) =
σ for T1 < u ≤ T2
where σ is the original stock price volatility. Thus the integrated volatility to
plug into the BS formula is
Z T2
σ̂ = σZ (u)du = σ · ∆
t

Since, by construction, Vt = St = s, and Zt = 1 we have the final answer

(remember the zero interest rate)

Π (t; X) = se−r∆ + s N [d1 ] − e−r∆ N [d2 ]

   
1 1 1 2
d1 (t) = √ ln + σ̂ (T2 − t) ,
σ̂ T2 − t K 2
p
d2 (t) = d1 (t) − σ̂ T2 − t.

Exercise 21.5 See the textbook, Chapter 7.

57
21.2.2 January 10, 2003

Exercise 21.6 With f (x, y) = xy we apply the Ito formula

1
df = fx dX + fy dY + fxx (dX)2 + fyy (dY )2 + fxy dXdY + fyx dY dX
2
to obtain
dZ = Z (α + β + σγ) dt + Z (σ + γ) dW
Standard technique shows that mt = E[Zt ] satisfies
dmt
= mt (α + β + σγ)
dt
so we obtain
E[Zt ] = x0 y0 e(α+β+σγ)t

Exercise 21.7 The AYLI is a T0 -claim X defined by

X = max [c(T0 ; K, T1 ), p(T0 ; K, T1 )]

where c(t; K, T ) and p(t; K, T ) denotes the prices at time t of a call (put) with
strike K and exercise date T on the given underlying. We now have put-call-
parity
p(T0 ; K, T1 ) = c(T0 ; K, T1 ) − ST0 + Ke−r(T1 −T0 )
so we obtain
h i
X = max c(T0 ; K, T1 ), c(T0 ; K, T1 ) − ST0 + Ke−r(T1 −T0 )
h i
= c(T0 ; K, T1 ) + max Ke−r(T1 −T0 ) − ST0 , 0

The last term is a put on S with exercise date T0 and strike price Ke−r(T1 −T0 ) .
The AYLI is thus equivalent to a sum of a put and a call (with different strikes
and exercise dates) and we get the price

Π(t) = c(t; K, T1 ) + p(t; Ke−r(T1 −T0 ) , T0 )

which, using put call parity and simplifying, can also be written

Π(t) = c(t; K, T1 ) + c(t; Ke−r(T1 −T0 ) , T0 ) − St + Ke−r(T1 −t)

n o n o
= St N [d1 ] − Ke−r(T1 −t) N [d2 ] + St N [d˜1 ] − Ke−r(T1 −T0 ) e−r(T0 −t) N [d˜2 ]
−St + Ke−r(T1 −t)
   
= St N [d1 ] + N [d˜1 ] − 1 − Ke−r(T1 −t) N [d2 ] + N [d˜2 ] − 1

58
where
    
1 St 1
d1 (t) = √ ln + r + σ 2 (T1 − t) ,
σ T1 − t K 2
p
d2 (t) = d1 (t) − σ̂ T1 − t
    
1 St 1 2
d˜1 (t) = √ ln + r + σ (T 0 − t) ,
σ T0 − t Ke−r(T1 −T0 ) 2
p
d˜2 (t) = d˜1 (t) − σ̂ T0 − t

Exercise 21.8 See textbook.

(a) Chapter 15
∂ ln p(t, T )
f (t, T ) = − ,
R ∂T T
− f (t,s)ds
p(t, T ) = e t

(b) Chapter 18
Z T
α(t, T ) = σ(t, T ) σ(t, s)ds
t

Exercise 21.9 See the textbook, Chapter 6. Remember all institutional as-
sumptions about no bid-ask spread, no tax effects, no transaction costs, liquid
market for the stock and option etc.

Exercise 21.10 The T -claim X to be priced is easily seen to be given by

X = S1 (T ) max {S2 (T ) − K, 0}

This is an obvious case of using S1 as numeraire, with martingale measure

denoted by Q1 and expectation denoted by E 1 . We obtain

Π (t; X) = S1 (t)E 1 [ max {S2 (T ) − K, 0}| Ft ]

From general theory we know that the Q1 dynamics of S2 are

dS2 (t) = (r + σ1 σ2 ) S2 (t)dt + S2 (t)σ2 dW 1 (t)

where σ1 σ2 denotes inner product and W 1 is Q1 -Wiener. Solving this GBM it

is easy to see that
S2 (T ) = e(r+σ1 σ2 )(T −t) S2 (t).

59
The expectation above is thus the price of a call on S with volatility kσ2 k strike
price K in a world with zero short rate. The total price is thus given by
 

Π (t; X) = S1 (t) · c e(r+σ1 σ2 )(T −t) S2 (t), t, K, kσ2 k, |{z}

0 , T
short rate

where, with obvious notation, c(s, t, K, σ, r, T ) as the Black-Scholes call option

formula.

21.2.3 March 18, 2002

Exercise 21.11 The solution F to the PDE is given by

 RT 
k(s,Xs )ds
F (t, x) = Et,x e t Φ(XT )

where

dXt = µ(s, Xs )ds + σ(s, Xs )dWs ,

Xt = x.

This is most easily proved by defining X as above and showing that the process
Rs
k(u,Xu )du
Zs = e t F (s, Xs )

is a Q-martingale, and, thus

 RT 
k(s,Xs )ds
Zt = EtQ [ZT ] ⇔ F (t, x) = Et,x e t Φ(XT ) .

Alternatively, it can be solved like Exercise 4.10 was solved in the Seminars
(check solutions given).

Exercise 21.12 The term structure equation has the form

1
FtT + {µ − λσ} FrT + σ 2 Frr
T
− rF T = 0,
2
F T (T, T ) = 1

See Chapter 16 in the textbook for details and discussion.

Exercise 21.13

60
 (a) The stock price of a dividend paying stock should be lower than the corre-
sponding price for a non dividend paying stock (see Chapter 11 for details).
Thus the price of call option should also be lower, recall that the delta of
a call is always positive (chapter 8).

(b) The general formula is (see Ch. 11 and extra exercise on Seminar 4)

F δ (t, s) = F 0 (t, se−δ(T −t) )

Applied to the standard BS formula we obain

F (t, s) = se−δ(T −t) N [d1 (t, s)] − e−r(T −t) KN [d2 (t, s)] .

     
1 s 1
d1 (t, s) = √ ln + r − δ + σ 2 (T − t) ,
σ T −t K 2
√
d2 (t, s) = d1 (t, s) − σ T − t.

Exercise 21.14 See Chapter 12 in the textbook for the arguments. The final
answer is

F (t, x) = xe−rf (T −t) N [d1 (t, x)] − e−rd (T −t) KN [d2 (t, x)] .

     
1 x 1
d1 (t, x) = √ ln + rd − rf + σ 2 (T − t) ,
σ T −t K 2
√
d2 (t, x) = d1 (t, x) − σ T − t.

Exercise 21.15 The same as Exercise 4 in the exam from March 2003.

61