Can Bitcoin, and other cryptocurrencies, be modeled effectively with a Markov Switching approach?
Written by Thomas Maupin
Thesis: Background A new era began at the end of 2008. The first cryptocurrency domain Bitcoin.org was launched (Mar-shall, 2019). Soon after, Satoshi Nakamoto released a white paper called "Peer-to-Peer Electronic Cash System" on October 31, 2008. The paper outlines the basis of the Bitcoin protocol and proposes a pure peer-to-peer currency that does not need to pass through or be influenced by anyone. Financial institutions (Satoshi Nakamoto, 2008). January 3, 2009 marked our history, because the first Bitcoin block, called the genesis block, was mined and rewarded with 50 Bitcoins ("How It All Started", 2019). Later that year, the Bitcoin software was released and the first Bitcoin transaction occurred. In October 2009, the first price of Bitcoin was 1 USD = 1309.03 BTC (Marshall, 2019). With a price, Bitcoin has finally reached its potential to become a currency. Businesses and individuals can trade Bitcoin, thus becoming the peer-to-peer electronic cash system that Nakomoto hopes. Fast forward to March 2010, and the first cryptocurrency transaction took place on the now-defunct website bitcoinmarket.com ("How It All Started", 2019). At that moment, cryptocurrency developed into a trading asset. As early as 2010, peers could mine, buy and exchange Bitcoin. The cryptocurrency competition started in mid-2013, because the crypto market consists of 10 digital assets, the main competitors Litecoin and Ripple ("How it all started", 2019). In 2015, the third main competitor, Ethereum appeared In the cryptocurrency market (Marshall, 2019). The three competitors and Bitcoin will be the cryptocurrencies of interest for this article. The decentralization of cryptocurrency is one of the main factors of its popularity (Marr, 2017). The lack of supervision allows peers to conduct anonymous exchanges between different countries and/or continents at a lower cost. One can define it as a means to achieve freedom of exchange. Since 2015, after people realized the freedom provided by cryptocurrency exchanges and the freedom to choose different types of cryptocurrencies, the crypto market has experienced many ups and downs. In May 2017, the price of one Bitcoin was $2,000 (Coinmarketcap, 2019). In December 2017, the price of a Bitcoin was close to 20,000 USD (Coinmarketcap, 2019). Within a few months, the price of Bitcoin skyrocketed, leading to a huge increase in popularity and conspiracy. At that moment, Bitcoin and other cryptocurrencies became the center of media attention, and there is now more and more research on this matter. 1.2 Literature review There have been many papers and articles that have studied the driving factors behind Bitcoin's price formation and volatility. Many of them were released in 2014, and some of them may be due to the sharp rise in Bitcoin prices at the end of 2013 and the subsequent international media reports. The price of Bitcoin has experienced slight price changes during its ten-year life cycle and has grown rapidly in 2017, rising to $19,535 throughout the year, and then falling to $13,000 in eight days (Coinmarketcap, 2019). The daily fluctuation of the price of Bitcoin (USD) is usually around 10%, and sometimes even reaches as high as 30%. Compared with the volatility experienced by major currencies such as the US dollar, the euro, and the Japanese yen in the foreign exchange market, the price of Bitcoin has experienced up to 30 times the volatility (Baur and Dimp, 2017). Rapid price changes have caused concerns among investors because it is difficult to predict the future value of Bitcoin. On the other hand, activist investors thrive in volatility because they are less risk-averse, do not care about stability, and do not have the idea of losing wealth (Aziz, 2019). In this paper, volatility refers to the fluctuation of prices over time. Unfortunately for them, Bitcoin continued to suffer severe price drops throughout 2018 (Coinmarketcap, 2019). Speaking of 2019, one can observe that Bitcoin seems to fluctuate between US$3,000 and US$4,000, and the price may rise. Uncertainty about how Bitcoin and other cryptocurrencies will develop drives the interest of this article. Can the history of cryptocurrencies provide any insights on how they have evolved over time? The fact that the price of Bitcoin is less than $1 and more than $19,000 (Coinmarketcap, 2019) makes us assume that standard time series models will not be able to be price correct. The gap between price and rapid volatility raises concerns about modeling and forecasting. How can statistical models have such historical time series? To overcome this problem, one can think of Bitcoin prices as different time series with different time intervals, linked to each other. Therefore, the Markov switching method was thought of. It is used to create a collection of time series related to transition probabilities. This means that every time series has a probability of occurrence and a probability of switching to another time series. ETH), Ripple (RPX) and Litecoin (LTC). The reason is that we not only want to understand the price movement of Bitcoin, but also how Bitcoin moves with other cryptocurrencies. The idea of the Markov conversion method on the first 4 cryptocurrencies brings us to the following question formula: Can Bitcoin and other cryptocurrencies be effectively modeled by the Markov conversion method? 1.4 Purpose When using financial time series analysis and modeling, we receive correlation coefficients that enable us to understand how a certain time series changes over time. When we analyze Bitcoin alone, we will be able to see how a certain price today depends on the price of the past. Read Less