Spectral calibration of exponential Lévy Models [1]



Spectral calibration of exponential Levy models ??

(c) With an estimate ψ of ψ at hand, we obtain estimators for the para-
metric part (
σ2 , γ, λ) by an averaging procedure taking into account
the polynomial structure in (11). Upon fixing the spectral cut-off value
U > 0, we set

ʌ

γ :=


U

2 :=/ Re(ψ(u))wU(u)

-U


du,


U

—σ2 + / Im( ψ( u )) wU ( u )d u,

-U


σ      σ2     ʌ fu ,. , ~, ʊ ,τ, , ,

λ := — + γ —    Re(ψ(u))wU(u) du,

2       -U


(14)

(15)

(16)


where the weight functions wσU , wγU and wλU satisfy

U2                U                 U

(17)


(18)


- —u- wσU ( u )d u =1,     uwY^ ( u )d u =1,     wU ( u )d u = 1;

U               -U            -U

U

U

0,      u2wλU (u) du = 0.

—U


wσU (u) du =

—U

The estimate of the coefficients can be understood as an orthogonal
projection estimate with respect to a weighted
L2-scalar product.

(d) Finally, we define the estimator for μ as the inverse Fourier transform
of the remainder:

μ(x) := F 1 h(ψ( ) + σ2-( i)2 iγ( i) + ʌ´ 1[u,u]( )i (χ). (19)

Then the identity (0) = ɪ — γ + λ shows that the estimated triplet
still satisfies the martingale condition (4).

3.2. Discussion of the method

First note that the computational complexity of the estimation procedure
is very low. Step (a) is a standard interpolation or regression estimation
procedure, which is well established and fast. The only time consuming
steps are the three integrations in step (c) and the (fast) Fourier transforms
in steps (a) and (d).

In step (a) a reasonable approximation of FO based on discrete data
must be found. Asymptotically, it suffices to use simple linear interpola-
tion because all regularisation takes place later in the spectral domain by
damping high frequencies. Depending on the observation design (
xj) and
the noise levels (
δj ), it may nevertheless pay off to invest more in obtaining
a good approximating function
O. When the distance j = xj — xj-1 be-
tween the transformed prices is rather large compared to the noise level
δj ,
the numerical approximation error prevails and higher order interpolation
schemes might significantly reduce the total error
O — O if O is smooth.
In Proposition 2 below, we only take advantage of the fact that
O is almost



More intriguing information

1. Rent-Seeking in Noxious Weed Regulations: Evidence from US States
2. The Complexity Era in Economics
3. Dementia Care Mapping and Patient-Centred Care in Australian residential homes: An economic evaluation of the CARE Study, CHERE Working Paper 2008/4
4. Retirement and the Poverty of the Elderly in Portugal
5. The migration of unskilled youth: Is there any wage gain?
6. The name is absent
7. The name is absent
8. Modellgestützte Politikberatung im Naturschutz: Zur „optimalen“ Flächennutzung in der Agrarlandschaft des Biosphärenreservates „Mittlere Elbe“
9. The name is absent
10. A Critical Examination of the Beliefs about Learning a Foreign Language at Primary School
11. Change in firm population and spatial variations: The case of Turkey
12. The name is absent
13. A Multimodal Framework for Computer Mediated Learning: The Reshaping of Curriculum Knowledge and Learning
14. Studying How E-Markets Evaluation Can Enhance Trust in Virtual Business Communities
15. The name is absent
16. Benchmarking Regional Innovation: A Comparison of Bavaria, Northern Ireland and the Republic of Ireland
17. Direct observations of the kinetics of migrating T-cells suggest active retention by endothelial cells with continual bidirectional migration
18. The name is absent
19. Forecasting Financial Crises and Contagion in Asia using Dynamic Factor Analysis
20. Education Responses to Climate Change and Quality: Two Parts of the Same Agenda?