Exercise Session 1

Fixed Income and Credit Risk

Question 1: Show by using heuristic arguments that for all t > s, Ws andh Wt − Ws are independent. i To do so
PN
you may use heuristic arguments as in class. Conclude what the value of E i=1 W ti−1 (W ti − W ti−1 ) is.

Correction Question 1: Set s < t and define ∆ ≡ T /N . Then, for N sufficiently large, there exists k and l such
that
k∆ ≤ s < (k + 1)∆, and l∆ ≤ t < (l + 1)∆.

Since Ws is approximated by
r k
T X
Xi ,
N i=1

and Wt − Ws is approximated by
r l
T X
Xi ,
N
i=k+1

no Xi figures in both sums, thus the sums are uncorrelated. Indeed,

r 
k r l
T X T X
Cov (Ws , Wt − Ws ) ∼ Cov  Xi , Xj 
N i=1 N
j=k+1
k l
T X X
= Cov (Xi , Xj )
N i=1 j=k+1
= 0.

Since Brownian increments are Gaussian, we conclude for independence. It follows that

XN N
X
E[ Wti−1 (Wtt − Wti−1 )] = E[Wti−1 (Wtt − Wti−1 )] = 0.
i=1 i=1

Question 2. Let Wt be a Brownian motion. At time t the Brownian has reached the value x.
q P
k
• Obtain the distribution of WT − Wt . Hint: Use the approximation that Wt ∼ N
T
i=1 Xi for Xi taking the
values +/-1 with equal probability?
Correction: As above, set t < T and define ∆ ≡ T /N . Then, for N sufficiently large, there exists k and l
such that
k∆ ≤ t < (k + 1)∆, and l∆ ≤ T < (l + 1)∆.

Then, WT − Wt is approximated by
r l
T X
Xi ,
N
i=k+1

1
 which is equal to r Pl 1
Pl
T (l − k) i=k+1 Xi
p l−k i=k+1 Xi − 0
· √ = ∆(l − k) · ,
N l−k √1
l−k

which is, by the CLT, distributed as N (0, T − t).

• What is the distribution of WT conditional on Wt = x? Denote the corresponding density of this distribution
p (WT , T ; Wt = x, t) and write it down explicitly.
Correction: Observe that
WT = Wt + (WT − Wt ),

where the first term is known and equals x, and WT − Wt is a normally distributed random variable with
zero mean and variance T − t. Then, the conditional density of WT is the density of a Normal distribution
with mean x and variance T − t:
2
1

WT −x
−1 √
p (WT , T ; Wt = x, t) = p e 2 T −t
.
2π(T − t)

• Show that this density follows a Partial Differential Equation of the type

∂2p ∂p
2
+k = 0,
∂x ∂t

where you are asked to determine k. Hint: You need to brutally compute the differentials.
Correction: With explicit differentiation, obtain:

∂p WT − x
=p ,
∂x T −t
2
∂2p

WT − x p
= p − ,
∂x2 T −t T −t
 2
∂p p WT − x p
=− + .
∂t 2 T −t 2(T − t)
Set k = 2 to obtain an identity
∂2p ∂p
+k = 0.
∂x2 ∂t

Question 3: Verify that the expression

N
X W12 W2
Wτi (Wti − Wti−1 ) = − 0
i=1
2 2
N N
1X
2 X
2
− Wti − Wti−1 + Wτi − Wti−1
2 i=1 i=1
N
X

+ (Wti − Wτi ) Wτi − Wti−1 .
i=1

given in the lecture notes is correct.

2
 Hint: develop the expression on the right hand side and simplify till you get the expression on the left side.

Correction Question 3: We develop several terms of the right hand side:

N N
1X
2 1X 2
− Wti − Wti−1 = − (W − 2Wti Wti−1 + Wt2i−1 ),
2 i=1 2 i=1 ti
N N
X
2 X
Wτi − Wti−1 = (Wτ2i − 2Wτi Wti−1 + Wt2i−1 ),
i=1 i=1
N
X N
X
(Wti Wτi − Wti Wti−1 − Wτ2i + Wτi Wti−1 ).


(Wti − Wτi ) Wτi − Wti−1 =
i=1 i=1

summation of these expressions yields

N N N N
X 1X 2 1X 2 X
W ti W τ i − Wti + Wti −1 − Wτi Wti−1
i=1
2 i=1 2 i=1 i=1
N
X
1 1
= Wτi Wti − Wti−1 − Wt2N + Wt20
i=1
2 2

Presently, remembering that t0 = 0 and that tN = T = 1 we conclude.