Beruflich Dokumente
Kultur Dokumente
Vadym Omelchenko
Normal Distribution
Laplace Distribution
t-Student Distribution
Cauchy Distribution
Logistic Distribution
Symmetric Stable Laws
Bivariate Normal
Distribution
0
N
0
1
,
Laplace bivariate
distribution
Definition of elliptical
distributions
T
X
(
X
,...,
X
)
The random vector
is said to
1
n
have an elliptical distribution with
(n 1)vector
(the
n matrix
n)
parameters
and
if its characteristic function can be
expressed as T
T
T
E exp(it X) exp(it ) ( t t ),
scalar
t some
(t1 , t 2 ,...,
t n ) function
T
for
and where
AAT
and are given by
E (it X) exp(it ) (t )
2
( t ) exp(it ) exp( t )
( t ) exp(it ) exp t 2
() exp
/2
/2
f X ( x)
cn
g n ( x )T 1 ( x )
(0)
Cov ( X )
Let X En ( , , g n ), let B be a
b R q. Then
b BX
q nmatrix and
Eq (b B , BB , g q )
T
X r Er ( 1 , 11 , g r )
En ( , , g.nThen
)
En r ( 2 , 22 , g n r )
Corollary. Let X
X nr
Convolutional Properties
If S n X 1 X 2 ... X n
and X ( X 1 , X 2 ,..., X n )
L ( X ) E n ( , , g n )
Then
L(S) E1 (e , e e, g1 )
T
Convolutional Properties
rA u
T
(k )
A A
T
The condition
n / 2 1
x
g n ( x ) dx
guarantees that
g n (.)
is a density generator.
(n / 2) 1
n / 2 1
cn
D
;
D
x
g n ( x)dx
n/2
(2 )
0
Examples of density
generators
g n (u ) (1 u ) ( n 1) / 2
Cauchy
Exponential
power
g n (u ) exp[ r u s ], r , s 0
Laplace
g n (u ) exp( u )
Logistic
g n (u )
Normal
g n (u ) exp(u / 2)
Student
exp(u )
[1 exp(u )]2
u
g n (u ) 1
m
( n m ) / 2
,m 0
an
integer
Construction of a density
function with infinite variance
1
g1 (u )
,
(1 u ) a
if
1 / 2 1
dx
g1 ( x) dx
a
0 (1 x )
dx
a 1 / 2
c1
a 1/ 2
f X ( x)
c1 1
c
g1
( x )2 1
2
1
a
( x )2
1
2
c
1
dx
EX x f X ( x)dx x 1
dx
const
a
2 a 1
x
1
( x )2
1
2
EX 2 x f X ( x)dx x 2
c1
if
a 3 / 2.
( x )2
2
dx const
dx
x
2a 2
Construction of a density
function with infinite variance
Financial Application
Expected Shortfall
TCE X ( xq ) E ( X X xq )
F ( xq ) 1 q
Expected Shortfall
2
1 xq
TCE X ( xq )
xq
Generalization of the
Previous Formula
g ( x)
dx
x
FZ (u ) c g 1 / 2 u 2 du ,
G ( x ) c g (u ) du
0
G ( ) ,
Define
~
G(x) G() - G(x)
Theorem 1
TCE X ( xq ) ,
2
Where is expressed as
~ 1 2
~ 1 2
G
zq
G
zq
2
2
FX ( xq )
FZ ( z q )
Examples
Logistic Distribution
1
1 ~ 1 2
1
(1 / 2 ) z q2
G
zq
1
2
2
1
(1 / 2 ) z q2
1
e
( zq )
1
1
2
2
1
1
1
(
1
/
2
)
z
2
2
q
e
( zq )
2
2
2
zq
1
1
FZ ( z q )
tanh
2
2
2
( zq )
(
z
)
q
TCE q
zq
1 tanh
2
En ( , , g n )
and
e (1,1,...,1)T is the
Suppose X
vector of ones with dimension n. Define
n
S n X 1 X 2 ... X n X k eT X
k 1
Theorem 2
j ,k
and
1 ~ 1 2
G z s ,q
S 2
S
,
FZ ( z s ,q )
with z s ,q ( sq S ) / S
and
e (1,1,...,1)T
Suppose X
is the
vector of ones with dimension n, and
n
S n X 1 X 2 ... X n X k eT X
k 1
k 1,2,..., n
k ,S
k ,S
k S
Skewed Elliptical
Distributions
All elliptical distributions belong to this
family.
All stable distributions belong to this family.
The density of the skewd Normal
Distribution has a form:
f ( x) 2 ( x) ( x)
Skewed
Stable
Distribution
s
Literatura
1. TAIL CONDITIONAL EXPECTATIONS
FOR ELLIPTICAL
DISTRIBUTIONS
Zinoviy M. Landsman* and Emiliano A.
Valdez
2. CAPM and Option Pricing with Elliptical
Distributions, Hamada M, Valdez.
3. Handbook of Heavy Tailed Distributions in
Finance, Eds S.T. Rachev