Modeling and Forecasting Multivariate Realized Volatility with Multivariate Fractional Brownian Motion
Dr. Chen ZHANG
Research Assistant Professor in Finance
FBA, UM
Date: 18 February 2025 (Tuesday)
Time: 13:00 to 14:00
Venue: FBA Lobby
Abstract
A multivariate fractional Brownian motion (mfBm) with component-wise Hurst exponents is used to model and forecast multivariate realized volatility. It is explored how the correlation coefficients interact with the Hurst exponents. A novel estimation method is proposed to estimate all parameters in the model. For time-reversible mfBm that is empirically reasonable for realized volatility, we develop asymptotic theory for all estimators. Optimal forecasting formulae are provided when the true data generating process is time-reversible mfBm and their properties are examined. When the time-reversible mfBm is used to forecast realized volatility out-of-sample, it is found that the model offers improvements over univariate fBm during the period when there are big differences among the estimated Hurst exponents, consistent with the prediction of our optimal forecast theory.
Speaker
Chen Zhang is a Research Assistant Professor of Finance at the University of Macau. He earned his PhD in Finance from Xiamen University and subsequently served as a Postdoctoral Researcher at Singapore Management University. His research primarily focuses on financial econometrics, with an emphasis on volatility models and yield curves. His work has appeared in several international academic journals, including the Journal of Econometrics, the Journal of Time Series Analysis, and Quantitative Finance.
All are welcome!
