This paper is concerned with testing the presence of ARCH within the ARCH-M model as the alternative hypothesis. Standard testing procedures are inapplicable since a nuisance parameter is unidentified under the null hypothesis. Nonetheless, the diagnostic tests for the presence of the conditional variance is very important since any misspecification in the conditional variance equation leads to inconsistent estimates of the conditional mean parameters. To resolve the problem of unidentified nuisance parameter, we apply Davies’ approach, and investigate its finite sample performance through a Monte Carlo study.
- ARCH ARCH-M Davies
- test LM test non-standard problem
ASJC Scopus subject areas
- Economics and Econometrics