The purpose of this paper is to examine the day-of-the-week effects of Bitcoin (BTC) markets on the exchange level from January 2014 to September 2018.
The in-depth study on the day-of-the-week effects is conducted by using data consisting of Bitcoin prices denominated in 20 fiat currencies from 23 Bitcoin trading exchanges through the method of rolling sample for calendar effect proposed by Zhang et al. (2017).
It is shown by the empirical results that different patterns of the day-of-the-week effects are observed on Bitcoin denominated in various fiat currencies by referring to the price data collected from exchanges. Furthermore, the patterns of the day-of-the-week effects are also available after adjusting Bitcoin prices denominated in domestic currencies into USD.
Because of the discontinuity of data for some daily return series, estimation with dynamic variance is not applicable. It is assumed that the error item follows normal distribution with constant variance.
The day-of-the-week effects are wide-spread in Bitcoin markets, and they are not mainly caused by movements of foreign exchange rates. Actually, empirical findings in this study provide evidence for inefficiency of Bitcoin markets.
