20 Pages Posted: 5 Dec 2003
Date Written: October 2003
In this paper a method is proposed for the testing and estimation of the GARCH effects. Estimation of the model is based on direct maximization of the log-likelihood function by numerical methods. Monte Carlo studies are conducted in order to evaluate the performace of various relevant designs. Also, we present Monte Carlo results showing that MLE estimators of the coefficient are less biased and more precise relative to their OLS counterpart in practically all the cases considered.
Keywords: Time-Series, estimation, simulation, Monte-Carlo methods
JEL Classification: C32, C51, C15, C22
Suggested Citation: Suggested Citation
By Robert Engle