Forecasting energy market volatility using GARCH models: Can multivariate models beat univariate models?

摘要

In this paper, we forecast energy market volatility using both univariate and multivariate GARCH-dass models. First, we forecast volatilities of individual assets and find that multivariate models display better performance than univariate models. Second, we forecast crack spread volatility and contrast the performance of multivariate models for two underlyings, with the alternative of univariate ones for crack spreads directly. Our evidence shows that univariate models allowing for asymmetric effects display the greatest accuracy. We also discuss the hedging strategy based on multivariate models and its implications for market participants.