(An Analysis of Bitcoin Prices with The Markov Regime Switching Model)

Abstract

Purpose – In the study, it is aimed to find the rising / gaining and falling / losing regimes, regime transition possibilities and duration of stay in the regime to determine the volatility dynamics of Bitcoin, which has the highest trade volume among crypto currencies. Design/methodology/approach – In the study, the rising / gaining and falling / losing regimes, regime transition possibilities and duration of stay in the regime were determined by the Markov Regime Switching Model, which allows both changes and regime transition probabilities to be calculated. Findings – In the study, for the study period it has been determined that the most suitable model for the Bitcoin return series is the three-regime MSIH (3) -AR (1) model which gives stronger results than the linear model. In the model consisting of three regimes, regime 1 with a negative coefficient indicates the contraction regime period, while regime 2 with positive coefficients indicates the transition and regime 3 the expansion regime period. It has been determined that while the Bitcoin returns series is in a regime, the probability of staying in the same regime for the next period is high, the possibility of transition from regime 1 to other regimes and from other regimes to regime 1 is low for the next period. Discussion – It has been determined that the regime persistence of Bitcoin returns is high for the period under study. In this context, if investors know the regime of Bitcoin returns in the current period, they have the opportunity to estimate the probability of staying in this regime and make their investment decision according to this information. However, considering that average duration of stay in the regimes is low, it is seen that active investors who keep Bitcoin in their portfolios will have the opportunity to increase the utility of their portfolios if they constantly follow the regime changes of this tool.