Implied-Volatility-Augmented GARCH Forecasting in Cryptocurrency and Traditional Asset Markets

This paper investigates whether implied volatility (IV) data can improve the accuracy of GARCH models in forecasting realised volatility (RV) for cryptocurrency in comparison with traditional assets. Specifically, we directly augment different variations of GARCH-type models (GARCH, EGARCH, GJR-GARCH) with IV indices. We then apply these IV-augmented GARCH models to bitcoin, ether, gold and crude oil for the first time to compare the degree of forecasting accuracy implied volatility provides and to test the superiority of the IV-augmented models over regression-type models. Our 1-day ahead forecasts confirm that the IV provides information beyond GARCH-estimated volatility when estimating realised volatility in all four assets. We further verify for the first time that the IV-augmented models, as opposed to their regression counterparts, provide the most accurate results for bitcoin, ether and oil. Taken together, we can conclude that cryptocurrency implied volatility, which is calculated by cryptocurrency option prices, provides as valuable information in estimating realised volatility as implied volatility does in the context of traditional asset markets. These results provide evidence that cryptocurrency options offer as unique information as that of traditional assets do in estimating volatility, paving the way for more effective risk management in the realm of cryptocurrency.