Realized Multivariate GARCH with Factors

Realized Multivariate GARCH with Factors

• Author
• Murilo Getlinger Coelho
• Abstract

• Forecasting second moments of asset returns is essential in portfolio selection. In a multivariate setting, the dimensionality of the problem and the precision of predictions are the main concerns. We propose a new methodology for forecasting covariance matrices joining two extant approaches in the literature: intraday data to enhance predictive ability and factors to reduce the dimensionality. We assume a multivariate realized GARCH model for the factors and a set of multivariate realized GARCH between each stock and the factors. We compare our methodology empirically with the standard literature by optimizing a portfolio on the S&P500 stocks universe.

• Keywords
• Financial Volatility, Realized GARCH, High Frequency Data, Multivariate Modeling, Correlation Matrix, Factors
• Subject Area
• Econometrics and Numerical Methods

Back Download