A1035
Title: Uncertainty and volatility: A Markov-switching GARCH-MIDAS approach
Authors: Yao Rao - The University of Liverpool (United Kingdom) [presenting]
Abstract: A flexible specification is proposed that encompasses virtually all GARCH-MIDAS models proposed in the literature. The specification accounts for asymmetry and regime switches in the volatility dynamics. In the empirical application, the relationship between S\&P 500 returns volatility and macroeconomic uncertainty is examined. It is shown that the model provides more accurate volatility forecasts than the nested GARCH-MIDAS models at forecast horizons of 1 day, two weeks, one month, two months and three months. Furthermore, the findings suggest that while high-frequency uncertainty indices are suitable for short-horizon forecasts, low-frequency uncertainty proxies provide better forecasts at longer horizons.
