USE OF CORRELATION AND REGRESSION ANALYSIS FOR THE CRYPTOASSETS’ PRICE MODELING IN ACCOUNTING AND CONTROL

The article investigates the use of correlation-regression analysis in cryptoassets' price modeling in accounting and control in a rapidly developing digital economy, its decentralization, the use of distributed registry technology (DTL) and blockchain technology. The author analyzed the dynamics of changes in the average capitalization of cryptocurrencies with the example of the most common cryptocurrency Bitcoin and other cryptocurrencies (Altcoins) during 2016–2020. In the absence of a unified definition of the term «cryptoassets», the author proposed a possible definition. A cryptoasset is a digital asset that uses cryptography, distributed ledger technology, and direct interaction between the two parties without the involvement of intermediaries. The author analyzed the necessity of determining the factors influencing the value of cryptoassets in order to predict their future price. In 2020, there were more than 5,784 different types of cryptoassets, including cryptocurrencies and tokens with a total market capitalization of over USD300,089 million. It is determined that factor analysis and cryptoassets' price forecasting should be carried out on the cryptoasset with the largest market capitalization — cryptocurrency Bitcoin, which represents more than 61% of the total capitalization of cryptoassets. The study confirmed the hypothesis that the value of the cryptocurrency Bitcoin is influenced by the indices of the world economy, the value of paired currencies, the value of precious metals, energy resources, shares of the largest US companies. The study of correlation-regression analysis involved 1,266 observations of each variable. The studied period was five years: from 07.08.2015 to 09.10.2020. The analysis provides an opportunity to rank the factors according to the strength of their influence on the cryptoasset's price. Therefore, the following factors had the greatest impact: the dynamics of the US financial services market, the Dow Jones industrial index, the NASDAQ general index, the value of Adobe Company's shares, the closing price of the cryptocurrency Ethereum (ETH). After the model optimization, the following three factors were left: the Dow Jones index; Ethereum cryptocurrency (ETH); Adobe stock price. The author prepared a multiple linear regression model. The resulting model can be used to predict the Bitcoin price in the short term. The average relative approximation error is 9%, which means that the model of a multiple regression as a whole adequately describes the relationship between the Bitcoin price and the selected factors in the short term.