Selecting the Order of an ARCH Model

Since the parameters of an autoregressive conditional heteroskedasticity (ARCH) process must be non-negative, inference on ARCH parameters can be improved by using inequality constrained estimation. In this paper, we extend this principle to the problem of ARCH lag order selection. We show that in the case of AIC, the appropriate adjustment to the penalty function has a simple form.

MoreLess

Year of publication:	1999
Authors:	Hughes, Anthony W. ; King, Maxwell L. ; Teng, Kwek Kian
Institutions:	School of Economics, University of Adelaide
Subject:	inequality constrained maximum likelihood \| one-sided AIC \| model selection

More details

Extent:	application/pdf
Series:	School of Economics Working Papers.
Type of publication:	Book / Working Paper
Language:	English
Notes:	Published in Economics Letters, 2004, vol. 83, issue 2, pp. 269-275 Number 1999-01 8 pages
Classification:	C22 - Time-Series Models ; C52 - Model Evaluation and Testing
Source:	RePEc - Research Papers in Economics

Persistent link: https://www.econbiz.de/10008462891