This paper studies the behaviour of Bitcoin returns at different sample frequencies. We consider high frequency returns starting from tick-by-tick price changes traded at the Bitstamp and Coinbase exchanges. We find evidence of a smooth intra-daily seasonality pattern, and an abnormal trade- and volatility intensity at Thursdays and Fridays. We find no predictability for Bitcoin returns at or above one day, though, we find predictability for sample frequencies up to 6 h. Predictability of Bitcoin returns is also found to be timeâvarying. We also study the behaviour of the realized volatility of Bitcoin. We document a remarkable high percentage of jumps above 80 % . We also find that realized volatility exhibits: (i) long memory; (ii) leverage effect; and (iii) no impact from lagged jumps. A forecast study shows that: (i) Bitcoin volatility has become more easy to predict after 2017; (ii) including a leverage component helps in volatility prediction; and (iii) prediction accuracy depends on the length of the forecast horizon.
Carlos TrucĂos, Aviral Kumar Tiwari, Faisal Alqahtani
Risk management is an important and helpful process for investors, hedge funds, traders and market makers. One of its key points is the appropriate estimation of risk measures which can improve the investment decisions and trading strategies. The high volatility of cryptocurrencies turns them a really risky investment and consequently, appropriate risk measures estimation is extremely necessary. In this article, we deal with the estimation of two widely used risk measures such as Value-at-Risk and Expected Shortfall in a cryptocurrency context. To face the presence of outliers and the correlation between cryptocurrencies, we propose a methodology based on vine copulas and robust volatility models. Our procedure is illustrated in a seven-dimensional equal-weight cryptocurrency portfolio and displays good performance.
Canh Phuc Nguyen, Nguyen Quang Binh, Thanh Dinh Su
The study examines the diversification capability of seven cryptocurrencies with the largest market size against risks from economic factors as oil price, gold price, interest rate, USD strength, and S&P500. Using the weekly data of Bitcoin, Litecoin, Ripple, Stellar, Monero, Dash, and Bytecoin in the period Aug/2014-Jun/2018, the study finds that there are structural breaks and ARCH disturbance in each cryptocurrency, suggesting a systematic risk within the cryptocurrency market. However, the causality between cryptocurrencies and economic factors is undirected. Interestingly, our findings show that cryptocurrencies are insignificant correlations with economic factors. The result implies that cryptocurrencies can not be assumed as financial assets to hedge systematic risks from economic factors.
Vasily Derbentsev, Natalia Datsenko, Olga Stepanenko, Vitaly Bezkorovainyi
This paper describes the construction of the short-term forecasting model of cryptocurrenciesâ prices using machine learning approach. The modified model of Binary Auto Regressive Tree (BART) is adapted from the standard models of regression trees and the data of the time series. BART combines the classic algorithm classification and regression trees (C&RT) and autoregressive models ARIMA. Using the BART model, we made a short-term forecast (from 5 to 30 days) for the 3 most capitalized cryptocurrencies: Bitcoin, Ethereum and Ripple. We found that the proposed approach was more accurate than the ARIMA-ARFIMA models in forecasting cryptocurrencies time series both in the periods of slow rising (falling) and in the periods of transition dynamics (change of trend).
Cryptocurrencies lack clear measures of fundamental values and are often associated with speculative bubbles. This paper introduces a new way of testing for speculative bubbles based on StockTwits sentiment, which is used as the transition variable in a smooth transition autoregression. The model allows for conditional heteroskedasticity and fat tails of the conditional distribution of the error term, and volatility may depend on the constructed sentiment index. We apply the model to the CRIX index, for which several bubble periods are identified. The detected locally explosive price dynamics, given the specified bubble regime controlled by a smooth transition function, are more akin to the notion of speculative bubble that is driven by exuberant sentiment. Furthermore, we find that volatility increases as the sentiment index decreases, which is analogous to the commonly called leverage effect.
Do the rich always get richer by investing in a cryptocurrency for which new coins are issued according to a proof-of-stake (PoS) protocol? We answer this question in the negative: Without trading, the investor shares in the cryptocurrency are martingales that converge to a well-defined limiting distribution and, hence, are stable in the long run. This result is robust to allowing trading when investors are risk neutral. Then, investors have no incentive to accumulate coins and gamble on the PoS protocol but weakly prefer not to trade. This paper was accepted by Kay Giesecke, finance.
ABSTRACT We find that three factorsâcryptocurrency market, size, and momentumâcapture the crossâsectional expected cryptocurrency returns. We consider a comprehensive list of priceâ and marketârelated return predictors in the stock market and construct their cryptocurrency counterparts. Ten cryptocurrency characteristics form successful longâshort strategies that generate sizable and statistically significant excess returns, and we show that all of these strategies are accounted for by the cryptocurrency threeâfactor model. Lastly, we examine potential underlying mechanisms of the cryptocurrency size and momentum effects.
Through the application of Diagonal BEKK and Asymmetric Diagonal BEKK methodologies to intra-day data for eight cryptocurrencies, this paper investigates not only conditional volatility dynamics of major cryptocurrencies, but also their volatility co-movements. We first provide evidence that all conditional variances are significantly affected by both previous squared errors and past conditional volatility. It is also shown that both methodologies indicate that cryptocurrency investors pay the most attention to news relating to Neo and the least attention to news relating to Dash, while shocks in OmiseGo persist the least and shocks in Bitcoin persist the most, although all of the considered cryptocurrencies possess high levels of persistence of volatility over time. We also demonstrate that the conditional covariances are significantly affected by both cross-products of past error terms and past conditional covariances, suggesting strong interdependencies between cryptocurrencies. It is also demonstrated that the Asymmetric Diagonal BEKK model is a superior choice of methodology, with our results suggesting significant asymmetric effects of positive and negative shocks in the conditional volatility of the price returns of all of our investigated cryptocurrencies, while the conditional covariances capture asymmetric effects of good and bad news accordingly. Finally, it is shown that time-varying conditional correlations exist, with our selected cryptocurrencies being strongly positively correlated, further highlighting interdependencies within cryptocurrency markets.
Wolfgang Karl HĂ€rdle, Campbell R. Harvey, Raphael Constantin Georg Reule
Abstract Cryptocurrency refers to a type of digital asset that uses distributed ledger, or blockchain, technology to enable a secure transaction. Although the technology is widely misunderstood, many central banks are considering launching their own national cryptocurrency. In contrast to most data in financial economics, detailed data on the history of every transaction in the cryptocurrency complex are freely available. Furthermore, empirically oriented research is only now beginning, presenting an extraordinary research opportunity for academia. We provide some insights into the mechanics of cryptocurrencies, describing summary statistics and focusing on potential future research avenues in financial economics.
Bitcoin has received a lot of attention from both investors and analysts, as it forms the highest market capitalization in the cryptocurrency market. This paper evaluates the volatility of Bitcoin returns using three GARCH models (sGARCH, iGARCH, and tGARCH). The new development allows for the modeling of volatility clustering effects, the leptokurtic and the skewed distribution in the return series of Bitcoin. Comparative to the Students't-distribution and the Generalized error distribution, the Normal Inverse Gaussian (NIG) distribution captured adequately the leptokurtic and skewness in all the GARCH models. The tGARCH model was the best model as it described the asymmetric occurrence of shocks in the Bitcoin market. That is, the response of investors to the same amount of good and bad news are distinct. From the empirical results, it can be concluded that tGARCH-NIG was the best model to estimate the volatility in the return series of Bitcoin. Generally, it would be optimal to use the NIG distribution in GARCH type models since time series of most cryptocurrency are leptokurtic.
Abstract This paper studies the volatility of Bitcoin and determines the importance of jumps and structural breaks in forecasting volatility. We show the importance of the decomposition of realized variance in the inâsample regressions using 18 competing heterogeneous autoregressive (HAR) models. In the outâofâsample setting, we find that the HARQâFâJ model is the superior model, indicating the importance of the temporal variation and squared jump components at different time horizons. We also show that HAR models with structural breaks outperform models without structural breaks across all forecasting horizons. Our results are robust to an alternative jump estimator and estimation method.
Bitcoin as a major cryptocurrency has come up as a shooting star of the 2017 and 2018 headlines. After exploding its price twenty times just in the twelve months of 2017, the tone has changed dramatically in 2018 after major price corrections and increasing concerns about its mining power consumption and overall sustainability. The dynamics and interaction between Bitcoin price and its mining costs have become of major interest. Here we show that these two quantities are tightly interconnected and they tend to a common long-term equilibrium. Mining costs adjust to the cryptocurrency price with the adjustment time of several months up to a year. Current developments suggest that we have arrived at a new era of Bitcoin mining where marginal (electricity) costs and mining efficiency play the prime role. Presented results open new avenues towards interpreting past and predicting future developments of the Bitcoin mining framework.
In todayâs era of big data, deep learning and artificial intelligence have formed the backbone for cryptocurrency portfolio optimization. Researchers have investigated various state of the art machine learning models to predict Bitcoin price and volatility. Machine learning models like recurrent neural network (RNN) and long short-term memory (LSTM) have been shown to perform better than traditional time series models in cryptocurrency price prediction. However, very few studies have applied sequence models with robust feature engineering to predict future pricing. In this study, we investigate a framework with a set of advanced machine learning forecasting methods with a fixed set of exogenous and endogenous factors to predict daily Bitcoin prices. We study and compare different approaches using the root mean squared error (RMSE). Experimental results show that the gated recurring unit (GRU) model with recurrent dropout performs better than popular existing models. We also show that simple trading strategies, when implemented with our proposed GRU model and with proper learning, can lead to financial gain.
Financial Technology (FinTech) is considered a taxonomy that describes a wide range of ICT (information and communications technology) associated with financial transactions and related operations. Improvement of service quality is the main issue addressed in this taxonomy, and there are a large number of emerging technologies including blockchain-based cryptocurrencies and smart contracts. Due to its innovative nature in accounting, blockchain can also be used in lots of other FinTech contexts where token models play an important role for financial engineering. This paper revisits some of the key concepts accumulated behind this trend, and shows a generalized understanding of the technology using an adapted stochastic process. With a focus on financial instruments using blockchain, research directions toward stable applications are identified with the help of a newly proposed stabilizer: interpretation function of token valuation. The idea of adapted stochastic process is essential for the stabilizer, too.
The development of new digital technologies in the areas of cryptography, distributed ledgers and mobile phones is affecting the way money is used for economic transactions. Electronic payments systems are rapidly replacing the use of cash. New powerful distributed ledger technologies, operated on a peer-to-peer decentralized basis is leading to the rapid expansion of digital money, with bitcoin being the most prominent digital currency (although there are more than one-thousand different crypto-currencies).
Roberto B. Corcino, Karl Patrick Casas, Allan Roy Elnar
In this paper, a mathematical model is constructed that would capture the pattern of the MSD of fluctuations of Bitcoin unit prices over time in daily basis by applying the method of White Noise Analysis. The raw data of 2805 ordered points (t,p) are used in the study, which are taken from coindesk.com, where p is the Bitcoin unit price at given time t.