Authors : Daniel Eni
Abstract: In this study, the performance of an Autocovariance Base Estimator (ABE) for GARCH models was studied, against that of the Maximum Likelihood Estimator (MLE) if the distribution assumption is wrongly specified as normal. We do this by first simulating time series data that fits GARCH model using the Log normal and t-distribution with degrees of freedom of 5, 10 and 15 as the true probability distribution but assumed normality in the process of parameter estimations. To keep track of consistency, we conduct and present the studies in sample sizes of 200, 500, 1000 and 1200. The two methods were then used to analyse the series under normality assumption. The result shows that the ABE method appears to be competitive in the situations considered.
Daniel Eni , 2010. Aperformance Ratings of an Autocovariance Base Estimator (ABE) in the Estimation of GARCH Model Parameters When the Normality Assumption is Invalid. Research Journal of Applied Sciences, 5: 108-111.