Investigation of supply-demand and price movements of an emerging technology product Bitcoin with Markov regime switching error correction model

This thesis analyses the relationship between Bitcoin price and supply – demand fundamentals of Bitcoin. In the thesis, daily data covering the 10.02.2017 – 26.03.2019 periods were used. Hypotheses were explained by using the Classical Unit Root, Johansen Cointegration and Markov Regime Switching Vector Error Correction Methods. Respectively, the supply and demand of Bitcoin are explained by Bitcoin Energy Consumption Index and Transaction Fee. As a control variable, Hashrate is included in the model. Bitcoin price is considered in BTC/USD. First, stationary situation of the variables was determined by Augmented Dickey Fuller and Phillips Perron Tests. As a results of these tests, it was found that the variables were stationary at I(1) level. The long term dynamics are characterized by three cointegrated vector with Johansen Cointegration Test. To determine best Markov Switching Model, the possible Markov Regime Change Error Correction Models in which the mean, constant term, autoregressive parameter and heteroscedasticity vary between regimes have been estimated. According to the information criteria, with three regimes and two lags MSIA(3) – VECM(2,2) model was chosen as the most suitable model. By using smoothed probabilities and transition probability matrix, the milestones were determined. Spending time and the average price of Bitcoin in each regimes was calculated. During the regime periods, the media coverage of Bitcoin was examined and the compatibility of the regimes with the Dow Theory was tested. In conclusion, although there are many external factors affecting the Bitcoin asset, it is seen that the MSIA (3) - VECM (2,2) model with three regimes and two lags represents market equilibrium and business cycles. In the first regime determined by the MSIA (3) - VECM (2,2) model, the market is in equilibrium. In this regime, while Bitcoin price and demand decrease, supply increases. In the second regime; while Bitcoin price and supply decrease, demand increases. "Bear Market" is effective in Regime 2. In the third regime, Bitcoin price and supply increase. "Bull Market" is effective in Regime 3. Keywords: Markov Switching, Stationary, Cointegration, Bitcoin, Supply, Demand, Price.