The impacts of outliers on different estimators for GARCH processes: an empirical study

The ARCH model introduced by Engle [1982] and extended to the GARCH model byBollerslev [1986] is able to capture some of the stylized facts of nancial data Rama [2001].The analysis of nancial time series is sensitive to outliers, due to the temporal dependencein the data. According to Carnero et al. [2007] the conditional heteroscedasticity of nancialdata could be explained through the presence of outliers in the data. Outliers can be explained,among other things, via errors in the transmission or deviations from the assumed model.This motivates the use for robust estimators for GARCH models. Section 2 summarizescurrent results for dierent estimators for GARCH processes.[...]