A1036
Title: What drives cryptocurrency returns? A sparse statistical jump model approach
Authors: Federico Cortese - University of Milano-Bicocca (Italy) [presenting]
Erik Lindstrom - Lund University (Sweden)
Petter Kolm - New York University (United States)
Abstract: The statistical sparse jump model, a recently developed, robust and interpretable regime-switching model, is used to analyze the factors driving the return dynamics of the largest cryptocurrencies. This method simultaneously incorporates feature selection, parameter estimation, and state classification. A wide range of candidate features is considered, including cryptocurrency, sentiment, and financial market-based time series that are known to influence cryptocurrency returns. The empirical analysis demonstrates that a three-state model provides a good representation of the cryptocurrency return dynamics. The latent states are interpreted as a bull, neutral, and bear market regimes, respectively. Through the data-driven feature selection approach, the significant factors are identified, and insignificant ones are excluded. The results indicate that within the candidate features, the first moments of returns, features indicating trends and reversal signals, market activity, and public attention are key drivers of crypto market dynamics.
