Advances in Risk Management

(Michael S) #1
296 NPV PROBABILITY DISTRIBUTION OF RISKY INVESTMENTS

Using the development of log(1+z), given by log(1+z)=z−z
2
2 +
z^3
3 −
z^4
4 +..., we obtain
the logarithm of the characteristic function in terms of its cumulants:


√∑αt ̃εt
∑α 2
t

=

∑n
t= 1

logφ ̃εt




αth
√∑
α^2 t




=

∑n

t= 1


i√αt

α^2 t

hK 1 −
1
2


√αt

α^2 t




2
h^2 K 2 −
i
3!


√αt

α^2 t




3
h^3 K 3

+
1
4!


√αt

α^2 t




4
h^4 K 4 +...




whereKjis the cumulant of orderj. By assumption, we have setK 1 =0etK 2 =1 whereas


it is obvious that


∑n
t= 1

(
√∑αt
α^2 t

) 2
=1. Hence:

√∑αt ̃εt
∑α 2
t

=

∑n
t= 1

logφε




αth
√∑
α^2 t



=−

h^2
2

i
3!

∑n
t= 1




αt
√∑
α^2 t




3
h^3 K 3

+
1
4!

∑n

t= 1


√αt

α^2 t




4
h^4 K 4 +...

which may be written as:


√∑αt ̃εt
∑α 2
t

=

∑n

t= 1

logφε




αth
√∑
α^2 t



=−

h^2
2

i
3!

∑n

t= 1

(
α^2

α^2 t

) 3 / 2
h^3 K 3

+
1
4!

∑n
t= 1

(
α^2 t

α^2 t

) 2
h^4 K 4 +...

and quite obviously:


nlim→∞

α^21
∑n
t= 1

α^2 t

=nlim→∞
(1+kc)−^2
∑n
t= 1

(1+kc)−^2 t

=
(1+kc)^2 − 1
(1+kc)^2
= 1 −
1
(1+kc)^2
= 0

wheneverkc =0.

Free download pdf