A comparison of alternative techniques for selecting an optimum ARCH model

Heather Mitchell, Michael D. McKenzie*

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

2 Citations (Scopus)

Abstract

Researchers have little to guide them when choosing an optimal model for use in auto regressive conditional heteroscedasticity modelling applications. Although the standard class of asymptotic model selection criteria may apply, some researchers have suggested that loss functions need to be developed, which are specific to each particular application. In this article, the relative merits of these two different techniques are considered. The results suggest that the model selection criteria provide superior results, although none of the techniques work well when the data are characterized by power and leverage effects.

Original languageEnglish
Pages (from-to)51-67
Number of pages17
JournalJournal of Statistical Computation and Simulation
Volume78
Issue number1
DOIs
Publication statusPublished - 2008

Keywords

  • ARCH
  • Model selection criteria

Fingerprint

Dive into the research topics of 'A comparison of alternative techniques for selecting an optimum ARCH model'. Together they form a unique fingerprint.

Cite this