Time-series Momentum in the Cryptocurrency Market

This thesis investigates the time-series momentum anomaly in the cryptocurrency market. Time-series momentum is the anomaly that the prior returns of an asset itself can be predictive for the future returns of the asset. To investigate the existence of the anomaly in the cryptocurrency market, three different time-series trading strategies are used: the simple TSMOM strategy, the filter strategy and the percentage price oscillator. The trading signals created by the different strategies are used to create equally-weighted momentum portfolios that are rebalanced daily. The returns of the equally-weighted momentum portfolios are compared to two custom-made benchmarks and the CCI30 index. This is done to create excess returns. Significant positive excess returns are found for 85% of the long-only portfolios. However, for the long-short portfolios no significant positive returns were found. Therefore, at least to some extent, evidence is provided that the momentum anomaly does exist in the cryptocurrency market.