Examining the relationships among cryptocurrencies using Google Trends
|Author:||Bc. Michael Heller|
|Year:||2019 - summer|
|Leaders:|| doc. PhDr. Ladislav Krištoufek Ph.D.
|Work type:|| Bachelors
|Awards and prizes:|
|Abstract:||The topic of our thesis is the examination of the relationships among cryptocurrencies using Google Trends. In our thesis, we concentrated on four cryptocurrencies, namely: Bitcoin, Litecoin, Ethereum Classic and Ethereum.
We obtained the data of daily opening prices, daily trading volumes and
daily Google Trends queries in order to examine the relationships among the
four cryptocurrencies. Applying the Vector autoregression model and Vector
error correction model, we constructed four models. The first model contains
only four time series of daily prices of cryptocurrencies. The second model is the
first model enriched by the respective four time series of Google Trends queries.
The third model contains the four time series of daily trading volumes of the four
cryptocurrencies. The fourth model is the third model enriched by the four time
series of Google Trends queries of respective cryptocurrencies. Then we applied
the Impulse response analysis and the Forecast error variance decomposition in
order to find some relationships among the variables. We found that there is
some correlation among prices, volumes and Google Trends queries containing
the names of the four cryptocurrencies. According to our results acquired by
the Forecast error variance decomposition, in all our models, Bitcoin has the
strongest explaining power of the variation of the variables.