Time series of financial returns are characterized by presenting heavy  tails, gain/loss asymmetry and volatility clusters. These characteristics make Generalized Orthogonal GARCH (GO-GARCH) models an excellent option for the modeling of such series, as it is a conditional volatility model for multivariate returns that can incorporate heavy tails and asymmetric returns quite naturally. In this report, we review the definition of GO-GARCH models and compare two di erent estimation strategies that are based on the Method of Moments and on Independent Component
Analysis using Brazilian market data.