Minimum density power divergence estimator for GARCH models

Abstract

In this paper, we study the robust estimation for the generalized autoregressive conditional heteroscedastic (GARCH) models with Gaussian errors. As a robust estimator, we consider a minimum density power divergence estimator (MDPDE) proposed by Basu et al. (Biometrika 85:549–559, 1998). It is shown that the MDPDE is strongly consistent and asymptotically normal. Our simulation study demonstrates that the MDPDE has robust properties in contrast to the maximum likelihood estimator. A real data analysis is performed for illustration.