CMStatistics 2018: Start Registration
View Submission - CFE
Title: Forecasting cryptocurrencies under model and parameter instability Authors:  Stefano Grassi - University of Rome 'Tor Vergata' (Italy) [presenting]
Leopoldo Catania - Aarhus BBS (Denmark)
Francesco Ravazzolo - Free University of Bozen-Bolzano (Italy)
Abstract: The predictability of cryptocurrencies time series is studied. We compare several alternative univariate and multivariate models in point and density forecasting of four of the most capitalized series: Bitcoin, Litecoin, Ripple, and Ethereum. We apply a set of crypto predictors and rely on dynamic model averaging to combine a large set of univariate dynamic linear models and several multivariate vector autoregressive models with different forms of time variation. We find statistically significant improvements in point forecasting when using combinations of univariate models and in density forecasting when relying on the selection of multivariate models. Both schemes deliver sizeable directional predictability.