[Forum SIS] Annuncio Seminario - Giuseppe Cavaliere and Anders Rahbek - 4 Giugno 2015

[Statlab- UNISA]

*************************************

Annuncio Seminario di Statistica

Dipartimento di Scienze Economiche e Statistiche - DiSES

Università degli Studi di Salerno

*************************************

Giovedì 4 giugno 2015, h. 11.00  Sala dei Consigli DiSES

Giuseppe Cavaliere (University of Bologna) and Anders Rahbek (University of Copenhagen)

 <http://www.dises.unisa.it/seminari/abstract/cavaliere_rahbek> Implementation and Theory of Bootstrap in Cointegrated Vector Autoregressive Models

ABSTRACT 

 

Part 1 (Giuseppe Cavaliere)

It is well known that the finite-sample properties of tests of hypotheses on the co-integration rank and on the co-integrating vectors in vector autoregressive (VAR) models can be quite poor, and that current solutions based on Bartlett-type corrections or bootstrap based on unrestricted parameter estimators are unsatisfactory, in particular in those cases where also asymptotic tests fail most severely. In this talk we solve this inference problem by initially showing that a bootstrap test on the cointegration rank where the null hypothesis is imposed on the bootstrap sample is asymptotically valid. That is, it has asymptotically correct size and it is also consistent under the alternative.

 

Part 2 (Anders Rahbek)

Contrary to bootstrap tests on the co-integration rank, establishing the validity of the bootstrap in the framework of hypotheses on the co-integrating vectors turns out to be a highly non-standard problem, which requires new theoretical arguments and the introduction of multivariate Ornstein-Uhlenbeck processes with random, reduced rank drift matrices. Again, we are able to show that a bootstrap test where the null hypothesis is imposed on the bootstrap sample not only does it have asymptotically correct size in the parameter region under the null but, surprisingly and in contrast to what is claimed in existing literature, it is consistent under the alternative. Finally, as documented by Monte Carlo simulation, the bootstrap tests also have excellent finite sample size and power properties, even for samples of moderate size.

 

Tutti gli interessati sono cordialmente invitati a partecipare.


Laboratorio di Ricerca e Formazione avanzata in Statistica (STATLAB) <http://www.dises.unisa.it/centri_laboratori/statlab/index> 

Dipartimento di Scienze Economiche e Statistiche - DiSES <http://www.dises.unisa.it/index>  

 

Università degli Studi di Salerno 

Via Giovanni Paolo II, 132

84084 Fisciano SA

Tel. +39 089 96 3132

 
-------------- parte successiva --------------
Un allegato HTML è stato rimosso...
URL: <http://www.stat.unipg.it/pipermail/sis/attachments/20150529/6d50b4fa/attachment.html>