de.mpg.escidoc.pubman.appbase.FacesBean

Post

 
 Vis
  Variance-Targeting in BEKK-ARCH Models
Item is

Ophav

 Ophav:
Søndergaard Pedersen, Rasmus 1, Forfatter
Rahbek, Anders 2, Vejleder
Tilknytninger:
1Økonomisk Institut, Det Samfundsvidenskabelige Fakultet, Københavns Universitet, København, Danmark, diskurs:7014              
2Det Samfundsvidenskabelige Fakultet, Københavns Universitet, København, Danmark, diskurs:7001              
skjul Ophav
Vis Ophav

Indhold

Ukontrollerede emneord: Multivariate ARCH, Variance-targeting, large-sample inference, geometric ergodicity
 Abstract: Objective The objective of this thesis is to investigate the large-sample properties of the variance-targeting estimator (VTE) for the BEKK-ARCH(1) model of Engle and Kroner (1995). The aim is to derive explicit, empirically relevant conditions for doing large-sample inference in the model. The theoretical results will be illustrated by simulations and in an empirical example where BEKK models are fitted on European stock-market return data by VTE.
skjul Indhold
Vis Indhold

Filer

Bemærkninger:
-
Tilgængelighed:
Offentlig
Mime-type / størrelse:
application/pdf / 523KB
Copyright dato:
2012-07-05
Copyright information:
De fulde rettigheder til dette materiale tilhører forfatteren.
skjul Filer
Vis Filer

Basal

Bogmærk denne post: https://diskurs.kb.dk/item/diskurs:31415:1
 Type: Speciale
Alternativ titel: Variance-targeting i BEKK-ARCH-modeller
skjul Basal
Vis Basal

Links

Vis Links

Detaljer

Sprog: English - eng
 Datoer: 2012-06-15
 Sider: -
 Publiceringsinfo: København : Københavns Universitet
 Indholdsfortegnelse: 1 Introduction 4
2 Notation and useful results 7
3 BEKK-ARCH(1) 8
3.1 The model and its reparametrization . 8
3.2 Properties of the BEKK-ARCH(1) process 9
3.3 Estimation 13
3.4 Pros and cons using VTE . 15
4 The large-sample properties of the VTE 17
4.1 Consistency . 17
4.2 Asymptotic normality . 19
5 Lemmas 23
5.1 Lemmas for the proof of consistency 25
5.2 Lemmas for the proof of asymptotic normality 32
6 Simulation study 46
6.1 Case 1: The DGP satis…es the su¢ cient conditions for asymptotic normality 46
6.2 Case 2: The DGP does not satisfy the su¢ cient conditions for asymptotic normality 48
6.3 Case 3: The DGP has E kXtk2 < 1, but E kXtk4 is not …nite . 50
7 Empirical illustration 52
8 Conclusion 57
A Appendix: Theorems 58
References 60
 Note: -
 Type: Speciale
skjul Detaljer
Vis Detaljer

Kilde

Vis Kilde