What is the real role of cryptocurrencies? This is the most frequent question which circulates among experts since the day Bitcoin appeared on the scene. Emprical evidence supplied by top economists’ studies starts to supply the first answers and shows how cryptos have a great potential to become part of any diversified portfolio. Regardless of discussions about their role as money, still questionable, it looks like they have clearly a role as an asset class in order to obtain a more performing, less risky portfolio.
Two researchers at Yale University, Yukun Liu and Aleh Tsyvinski, in their paper ‘Risks and Returns of Cryptocurrency’ (2018) investigated potential predictors for cryptocurrency returns that mirror those of traditional asset classes. Specifically, they constructed cryptocurrency momentum, proxies for average and negative investor attention, a proxy for price to-“dividend” ratio, realized volatility, and proxies for the supply conditions.
What they discovered looks revolutionary to asset managers . First, they found that “there is significant time-series cryptocurrency momentum at the daily and weekly frequencies for all three cryptocurrencies. For example, a one-standard-deviation increase in the current day’s Bitcoin return predicts a 0.33% increase in the daily return over the next day”.
Furthermore, grouping weekly returns by quintiles, they discovered that “the top quintiles outperform the bottom quintiles over the 1-4 week horizons. For example, at the 1-week horizon, the average return of the top quintile is 11.22% per week with the Sharpe ratio of 0.45 while the average return of the bottom quintile is 2.60% per week with the Sharpe ratio of 0.19. However, for Ethereum, the momentum effect is less significant than for Bitcoin and Ripple.”
Researchers discovered that “high investor attention predicts high future returns over 1-2 week horizons for Bitcoin, a 1-week horizon for Ripple, and 1-, 3-, and 6-week horizons for Ethereum. For example, a one-standard-deviation increase in the Google search for the word ‘Bitcoin’ yields a 2.3% increase in the 2-week ahead Bitcoin returns. At the 1-week horizon, the average return of the top quintile is 11.20% per week with the Sharpe ratio of 0.48 while the average return of the bottom quintile is 1.07% per week with the Sharpe ratio of 0.08.” Another proxy for investor attention they built is Twitter post counts. A one-standard-deviation increase in the Twitter post count for the word ‘Bitcoin’ yields a 2.50% increase in the 1-week ahead Bitcoin returns. Finally, they built a proxy for negative investor attention and demonstrated that “relatively high negative investor attention negatively predicts future Bitcoin returns”. They construct a ratio between Google searches for the phrase ‘Bitcoin hack’ and searches for the word ‘Bitcoin,’ and showed that “a one-standard-devation increase of the ratio leads to a 2.75% decrease in Bitcoin returns the following week.”
Third, they built a proxy for the the price-to-“dividend” ratios for Bitcoin and found that “it has no predictive power. Realized volatility also does not predict returns of Bitcoin and Ethereum but predicts Ripple returns at 4-, 5-, and 7-day ahead frequencies.”
Finally, they constructed proxies for the cost of mining to capture the supply factors, and found that “those have low loadings for Bitcoin and Ripple.”
The two authors main conclusion is clearcut: indeed cryptocurrencies are an asset class that can be assessed using simple finance tools. At the same time, cryptocurrencies comprise an asset class which is radically different from traditional asset classes. This is what make them suitable for being a component of any diversified portfolio.