Affordable Access

Publisher Website

Robustness of size of tests of autocorrelation and heteroscedasticity to nonnormality

Authors
Journal
Journal of Econometrics
0304-4076
Publisher
Elsevier
Publication Date
Volume
51
Identifiers
DOI: 10.1016/0304-4076(92)90027-o

Abstract

Abstract A comprehensive empirical examination is made of the sensitivity of tests of disturbance covariance in the linear regression model to nonnormal disturbance behaviour. Tests of autocorrelation appear to be quite robust, except for extreme nonnormality, but tests for heteroscedasticity are highly susceptible to kurtosis.

There are no comments yet on this publication. Be the first to share your thoughts.