ERRATA CORRIGE

(Last update: April 3, 2012)

PDE and Martingale Methods in Option Pricing
Andrea Pascucci
http://www.dm.unibo.it/pascucci/
Bocconi & Springer Series
ISBN: 978-88-470-1780-1
Chapter 2
p.53, +6: ...covers the whole R
>0
as N .
p.68, +12: ...measure P.
Chapter 3
p.131, -3: if T / then..
Chapter 4
p.149, Exercise 4.16: Under the hypotheses of Theorem 4.11, prove that for ( T
a
we have in general
E
__
b
a
u
t
dW
t
[ (
_
,=
_
b
a
E [u
t
[ (] dW
t
.
p.154, -1:
= E
__
t
s
u
2

d X
t
[ T
s
_
+X
2
s
= M
2
s
X
s
.
Chapter 6
p.206, -8: Proof. First of all, by Remark 6.8 ....
p.212, -2: Indeed we have u(t
0
, x) = 0 and
Chapter 7
p.222, 10:
d

V
(,)
t
=

S
t

t
dW
t
Chapter 8
p.262: estimates (8.13)-(8.14) of Theorem 8.10 hold true under the additional regularity Hypothesis 8.13.
In general, under Hypotheses 8.1 and 8.3, we have
[
x
i
(t, x; s, y)[
C

t s

+
(t s, x y),

x
i
x
j
(t, x; s, y)

+[
t
(t, x; s, y)[
C
t s

+
(t s, x y).
Chapter 9
1
p.306, -11: Formula (9.55) is related to the classical method of characteristics which can be used to solve
the initial value problem for general rst order (containing only rst order partial derivatives) PDEs
p.314, -2:
_

(t) = B(t)(t),
(0) = I
N
,
p.316, formula (9.75):
L =
1
2
N

i,j=1
c
ij
(t)
x
i
x
j
+
N

i=1
b
i
(t)
x
i
+
N

i=1
B
ij
(t)x
j

x
i
+
t
p.316, 7: The case of constant coecients...
Chapter 10
p.341, 4:

ii
=
d

j=1
_
A
ij
_
2
= 1.
Then, by Corollary 5.35,
W
i
t
=
d

j=1
A
ij

W
j
t
, 1 = 1, . . . , d,
is a standard 1-dimensional Brownian motion...
p.371, 9: Therefore, assuming the dynamics (10.92), we get
d
i
t
=
p
i
0
p
i1
0

i
L
i
t

i
t
dW
i,i
t
=
i
t

i
t

i
L
i
t
1 +
i
L
i
t
dW
i,i
t
.
p.371, formula (10.96):

i
t
=
i
t
N

k=i+1

ik

k
t

k
L
k
t
1 +
k
L
k
t
, i < N,
p.371, -2:
dL
i
t
= L
i
t
N

k=i+1

ik

i
t

k
t

k
L
k
t
1 +
k
L
k
t
dt +
i
t
L
i
t
dW
N,i
t
, i = 1, . . . , N 1,
p.375, 6:
= E
Q
_
e

T
0
rsds
(S
T
K)1
{S
T
K}
_
= I
1
I
2
p.375, 10: by the change of numeraire...
p.376, 14:
dQ
T
dQ
[
F
W
t
=
p(t, T)B
0
B
t
p(0, T)
.
p.380, 7: the SDE (10.111) has a unique strong solution for any [0, 1] and not only for
1
2
. A proof
of the fact that
u(t, s) = E
_
(S
t,s
T
K)
+
_
solves the Cauchy problem (10.112) can be found in
Janson, S. and Tysk, J. Feynman-Kac formulas for Black-Scholes-type operators. Bull. London Math.
Soc. 38 (2006), no. 2, 269282
2
p.382, 2:
F
t
=
e
r(Tt)
S
t
+K
2
.
p.383, formula (10.120):
_

Q
1
2

xx
Q = x
1

xx
Q+

2
x
2
(
2
1
+
2
)
2

xx
Q+ O(
3
), > 0, x >
K

,
Q(0, x) = x
+
, x >
K

.
p.383, -4:
_

t

Q
1
2

xx

Q = x
1

xx

Q+

2
x
2
(
2
1
+
2
)
2

xx

Q+ O(
3
), t > 0, x >
K

Q(0, x) = x
+
, x >
K

,
p.384, formula (10.127):

t
G(t, x) =
1
2
(t, x),
x
G(t, x) =
_
x

(t, y)dy,
Chapter 13
p.434, formula (13.3):
P (N
t
= n) = e
t
(t)
n
n!
, n N 0,
p.445, -8:
=

n1
P(N
t
= n)
n

k=1
E[
Z
k
(H)]
p.447, formula (13.26):
= tE[f(Z
1
)] = E [N
t
f (Z
1
)] .
p.448, +2:
E
__
t
s
_
R
d
f(x)J (d, dx) [ T
s
_
= E
_
_

Ns<nNt
f(Z
n
)
_
_
p.448, in the main formula

n1
should read

n0
p.451, formula (13.38):
t J
t
(H, f) :=
_
t
0
_
H
f(x)J(ds, dx) =

0<st
f(X
s
)1
H
(X
s
)
p.453, delete the second line of formula (13.51): thus formula (13.51) reads

X
,R
t
=
_
t
0
_
|x|<R
x(J(ds, dx) (dx)ds)
p.454, formula (13.52):

R
=
S
+
_
S|x|<R
x(dx).
p.454, 16:

R
t +
_
t
0
_
S|x|<R
x

J(ds, dx) =
S
t +
_
S|x|<R
x(dx),
p.454, formula (13.53):
_
|x|<1
[x[(dx) < ,
3
p.455, formula (13.54):

0
=
1

_
|x|<1
x(dx) < ,
p.455, formula (13.57):

:= lim
R+

R
=
S
+
_
|x|S
x(dx).
p.455, delete the second line of formula (13.58): formula (13.58) reads
X
t
=

t +B
t
+
_
t
0
_
R
d
x

J(ds, dx)
p.456, +5: If Z
1
is integrable then conditions (13.53) and (13.56) are satised...
p.456, +9:

1
= +
_
|x|<1
x(dx);
p.456, +10: the -triplet of X is ( +E [Z
1
] , (, ) (see also (13.73)).
p.458, +3: Proof. By the Levy-Ito decomposition X
t
is equal a.s. to the limit of the sum...
p.458, formula (13.63):

R
=
1
+
_
1|x|<R
x(dx).
p.460, formula (13.67):

0
=
1

_
|x|<1
x(dx).
p.460, formula (13.70):
c
2
(X
t
) = t
_

2
+
_
R
x
2
(dx)
_
,
p.461, formula (13.71):
c
n
(X
t
) = t
_
R
x
n
(dx), n 3.
p.465, +3:

Xt
(x) = e
t

n=0
(t)
n
n!
_
2(
2
t +n
2
)
e

(xtnm)
2
2(
2
t+n
2
)
.
p.481, formula (13.91):
E
Q
[S
T
] = E
Q
_
S
0
e
X
T

<
Chapter 14
p.499, +4: Thus in this case (i.e. when T
1

n
for any n) the Riemann-Stieltjes integral is well-dened...
p.524, +4: then we can let R go to zero in (14.45)
p.530, +16: Regularity of elliptic ( > 0)...
Chapter 15
p.571, 5: Setting
0
= u
0
and

= u, the rst two cumulants are given by

p.572, -2: Here S
0
= 130, K = 100 and T = 5: the reference value, computed with N = 1000, is 4.424162989.
4
Chapter 16
p.593, +13:

S
T
= (W
T
2T)S
T
, D
s
S
T
= S
T
.
Acknowledgements. Thanks are due to Emmanuel Gobet, Laura Monti, Stefano Pagliarani, Candia Riga,
Stefano Biagi for spotting some of the above errors.
5