Research NBER

Search in Google helps to predict changes in bitcoin prices

In a recent study published by the National Bureau of Economic Research (NBER), it is assumed that the crypto-currency markets, unlike traditional financial markets, vary depending on the type of attention they receive. Unlike other traditional financial assets, crypto currencies do not react and do not behave in accordance with the same set of market factors that are typical for traditional financial instruments. Instead, they are more closely linked to the “specific factors of the Crypto-currency,” according to a report released this week.

These factors include investor attention and market dynamics, and describe NBER as “time series of currency and currency dynamics at daily and weekly frequencies.”

The authors of the article, economists at Yale University Yukun Liu and Oleg Tsyvinsky, believe that, contrary to public opinion, “the markets do not consider crypto-currencies in the same way as standard asset classes.”

As a source of market data, the article used bitcoin, ether and ripple tracking services on CoinDesk. Using a series of price data for multi-year time intervals, the document compared actual and projected revenues using a standard financial pricing model known as CAPM.

Liu and Tsyvinsky compared revenues from the crypto currency with traditional currencies, such as the euro, and metals, such as gold, as well as macroeconomic factors — for example, consumption growth. All this led to unrepresentative results suggesting that the true model lies in another area that Liu and Tsyvinsky define as an indicator of return on investment for a day or a week.

Essentially, one, three, five, or six days’ price growth can be predicted by one day return of investment, while a weekly indicator can predict market movements for one, two, three or four weeks.

It is noteworthy that this research includes data on consumer activity in the Google search service and social networks, in particular, on Twitter. The researchers found that an increase in the standard deviation in the search for keywords such as bitcoin, predicted a slight increase in the price of the token in the following weeks.

On average, an increase in the average quadratic deviation in the search for keywords leads to an increase in the price by 2.75%, the report says. Similarly, an increase in this deviation in the number of entries in Twitter led to an increase in the price of bitcoin by 2.5%. On the other hand, an increase in this indicator in the case of “bitcoin-related hacking” predicted a slight decrease in the price of the crypto currency.

#NBER # bitcoin #Google # forecasting