Cryptocurrency factor momentum

Abstract

Is there a momentum effect in cryptocurrency anomalies? To answer this, we analyze data from over 3900 coins spanning the years 2014 to 2022 and replicate 34 anomalies in the cross-section of cryptocurrency returns. We document a discernible pattern in factor premia: past winners consistently outperform losers. The effect persists across subperiods, withstands various methodological approaches, and its magnitude parallels that of its stock market counterpart. However, the autocorrelation in factor returns is not widespread and primarily stems from size and volatility anomalies. Additionally, unlike in stocks, cryptocurrency factor momentum originates from price momentum, which subsequently transfers to the factor level.

Keywords:

• Factor momentum
• Cryptocurrency anomalies
• The cross-section of cryptocurrency returns
• Return predictability

JEL Classifications:

• G12
• G14
• G11
• G10

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 Note that the mean return of the cross-sectional factor momentum strategy is half of the difference between its long and short legs.

2 See Table IA.I on page 3 in the Internet Appendix to Ehsani and Linnainmaa (2022).

3 Specifically, at the beginning of each portfolio holding period, we include only those cryptocurrencies with an Amihud (2002) measure below the 50% and 25% percentile (see table 2 for the variable description). Note that the cryptocurrency market is extremely skewed as a few large cryptocurrencies account for the majority of the aggregate market capitalization. By only looking at the 50% and 25% most liquid cryptocurrencies, we still cover, on average, 98.6% and 97.1% of the aggregate market capitalization, respectively. Therefore, this restricted sample is of high relevance for real-world cryptocurrency trading.

Additional information

Funding

This work was supported by Narodowe Centrum Nauki [2021/41/B/HS4/02443].