Volatility Forecasting in Bitcoin Market : A New Proposed Measure Based on Vs-Acarr Approach

In this paper, we propose a new Volatility-Spillover-Asymmetric Conditional-Auto-Regressive-Range (VS-ACARR) approach to model/forecast Bitcoin’s volatility (Price-range). Traditionally, CARR models are uni-variate but we introduce volatility spillover from another series (Crude oil) in the forecasting model construction. The proposed VS-ACARR model takes into account both the spillover as well as asymmetry (leverage effect). The empirical application to oil price (WTI) and Bitcoin price-range shows the existence of strong volatility spillover from crude oil to Bitcoin market and weak leverage effect in Bitcoin market. Moreover, the VS-ACARR model yields the higher forecasting accuracy compared with the GARCH, the CARR and the VS-CARR models in terms of out-of-sample forecast performance, implying that accounting for volatility spillover and volatility asymmetry can significantly improve the forecasting accuracy of Bitcoin price-range (volatility). The superior forecasting performance of the VS-ACARR model is lasting in case of alternative forecasting windows and a bunch of other robustness tests. This finding highlights the significance of accommodating spillover information from crude oil and volatility asymmetry in forecasting Bitcoin’s volatility