Volatility forecasting with bivariate multifractal models

• DOI: 10.1002/for.2619
• Publication Date: 2019-07-16T07:58:21Z
• Authors: Ruipeng Liu; Riza Demirer; Rangan Gupta; Mark Wohar
• Keywords: Multifractal system; Univariate; Stylized fact

Abstract This paper examines volatility linkages and forecasting for stock foreign exchange markets from a novel perspective by utilizing bivariate Markov‐switching multifractal model that accounts possible interactions between markets. Examining daily data major advanced emerging nations, we show generalized autoregressive conditional heteroskedasticity models generally offer superior forecasts short horizons, particularly returns in Multifractal models, on the other hand, significant improvements longer consistently across most Finally, provides compared to univariate alternative more currency returns, while its benefits are limited case of