Predicting Bitcoin Returns with Momentum Effect and Investor Attention Effect

Yale economists have proposed a new model for predicting bitcoin’s future valuation based on historical price data and Google search queries. According to the researchers, momentum and investor attention are the two key factors that can enable market participants to accurately predict future returns.

Predicting Bitcoin

In a newly published study called “Risks and Returns of Cryptocurrency,” economists Aleh Tsyvinski and Yukun Liu claim the Momentum Effect and Investor Attention Effect can accurately predict the future valuation of leading cryptocurrencies like bitcoin, Ethereum and Ripple XRP.

The Momentum Effect

The Momentum Effect essentially states that cryptocurrency prices are more likely to continue rising following a significant breakout. For example, a large rally in bitcoin one week is likely to generate continued growth the following week. In this vein, momentum applies to cryptocurrencies in the same way it does to stocks, bonds and currencies, though the size of the rallies are usually much larger tha conventional assets

“Momentum is actually something simple,” Tsyvinski said in an interview with CNBC. “If things go up, they continue to go up on average, and if things go down, they continue to go down.”

The researchers believe that the best strategy for making money in crypto is to buy an asset after its price has already spiked and sell it just seven days after purchase. Using this strategy, a trader can still make an average of 11% on bitcoin even if they bought it following a 20% increase.

Tsyvinski explained that the Momentum Effect was stronger for bitcoin than for Ethereum or XPR, although still statistically significant for the latter two.

The Investor Attention Effect

The level of interest and hype around cryptocurrencies also plays a significant role in predicting price movements, the researchers claim. The Investor Attention Effect measures Google and social media search trends for terms like “bitcoin” “Ethereum,” and “cryptocurrency.” The higher the search results, the greater the chances of rising prices.

“For weekly returns, the Google search proxy statistically significantly predicts 1-week and 2-week ahead returns,” the report says.

This is not unlike what Hacked has reported several times before: Google search trends for crypto keywords are positively correlated with rising values. In June, we reported that Google searches for “bitcoin” had fallen to nine-month lows, a sign that first-time buyers were no longer flooding the market.

Google search trends gave “bitcoin” a perfect score of 100 back in December, around the time that BTC was trading at all time highs. Our view at the time was the following:

“Bitcoin’s bull market was largely predicated on the arrival of new traders buying cryptocurrency for the first time… Without new first-time buyers entering the market, bitcoin could face a prolonged lull period characterized by sideways movement and false breakout patterns.”

That outlook has largely panned out in the midst of multiple peaks and troughs.

Disclaimer: The author owns bitcoin, Ethereum and other cryptocurrencies. He holds investment positions in the coins, but does not engage in short-term or day-trading.

Featured image courtesy of Shutterstock.

Author:
Chief Editor to Hacked.com and Contributor to CCN.com, Sam Bourgi has spent the past nine years focused on economics, markets and cryptocurrencies. His work has been featured in and cited by some of the world's leading newscasts, including Barron's, CBOE and Forbes. Avid crypto watchers and those with a libertarian persuasion can follow him on twitter at @hsbourgi