Autoregressive Conditional Heteroskedasticity in the Presence of Serial Correlation

by Stan Hurn and David McDonald .

Abstract: Recently Bera et al. proposed a test for time-varying variance of regression errors, known as autoregressive conditional heteroskedasticity (ARCH), in the presence of serial correlation. We examine the empirical size and power of this and three other tests. The primary purpose is to provide information to applied researchers who might wish to choose selectively from a range of diagnostic tests when ARCH and/or serial correlation are likely to feature in the data.

Key Words: size, power, autoregressive conditional heteroskedasticity, serial correlation

Authors:
Stan Hurn, shurn@cupid.ecom.unimelb.edu.au
David McDonald, david.mcdonald@ml.csiro.au

Editor: Don Edwards , edwards@math.scarolina.edu

READING THE ARTICLE: You can read the article in portable document (.pdf) format (89557 bytes.)

NOTE: The content of this article is the intellectual property of the authors, who retains all rights to future publication.

This page has been accessed 2365 times since April 1, 2007.


Return to the InterStat Home Page.