The role of uncertainty index in forecasting volatility of Bitcoin: Fresh evidence from GARCH-MIDAS approach

Abstract

We shed light on the role of Economic Policy Uncertainty (EPU) and Cryptocurrency Uncertainty (UCRY) indices in forecasting Bitcoin volatility. The empirical results of in-sample estimations demonstrate that the global EPU index and UCRY indices exhibit significantly negative and positive effects on long-term Bitcoin volatility, respectively. Moreover, the out-of-sample validation reveals that One-Side Asymmetric GARCH-MIDAS with UCRY price index is the best-performing model and forecasting models incorporating the UCRY indices significantly outperform models with global and national EPUs in out-of-sample forecasting. Considering its scarce application, UCRY indices become a promising data source in guiding Bitcoin trading behaviors.