Yang Zhou, Chi Xie, GangâJin Wang, Jue Gong · 5 authors
Abstract Cryptocurrency is a remarkable financial innovation that has affected the financial system in fundamental ways. Its increasingly complex interactions with the conventional financial market make precisely forecasting its volatility increasingly challenging. To this end, we propose a novel framework based on the evolving multiscale graph neural network (EMGNN). Specifically, we embed a graph that depicts the interactions between the cryptocurrency and conventional financial markets into the predictive process. Furthermore, we employ hierarchical evolving graph structure learners to model the dynamic and scale-specific interactions. We also evaluate our frameworkâs robustness and discuss its interpretability by extracting the learned graph structure. The empirical results show that (i) cryptocurrency volatility is not isolated from the conventional market, and the embedded graph can provide effective information for prediction; (ii) the EMGNN-based forecasting framework generally yields outstanding and robust performance in terms of multiple volatility estimators, cryptocurrency samples, forecasting horizons, and evaluation criteria; and (iii) the graph structure in the predictive process varies over time and scales and is well captured by our framework. Overall, our work provides new insights into risk management for market participants and into policy formulation for authorities.
This paper aims to investigate investorsâ prospects in adding value to their portfolios by considering investorsâ behavioural score (Cumulative Prospect Theory (CPT) score) and a clustering technique in the selection of assets. The universe of assets constitutes 63 cryptocurrencies sourced from Bloomberg from Jan 01, 2020, to July 31, 2022. The study period was segmented into two distinct and mutually exclusive periods, namely COVID-19, and post-COVID-19. Nine portfolios were constructed of which six were based on the CPT and the remaining on the K means Clustering technique. Using the copula-based Differential Evolution (DE) algorithm for the optimisation, the results show that portfolios consisting of assets with extremely high CPT scores were preferred during the post-COVID-19 and full sample periods, except for portfolios comprising assets with extremely low CPT scores during the COVID-19 period. The most optimised portfolio was composed of classified assets with extremely high CPT scores in the post-COVID-19 period. These findings provide intuitive and coherent investment strategies to guide investors in the cryptocurrency market.
Our research investigates the predictive performance and robustness of machine learning classification models and technical indicators for algorithmic trading in the volatile cryptocurrency market. The main aim is to identify reliable approaches for informed decision-making and profitable strategy development. With the increasing global adoption of cryptocurrency, robust trading models are essential for navigating its unique challenges and seizing investment opportunities. This study contributes to the field by offering a novel comparison of models, including logistic regression, random forest, and gradient boosting, under different data configurations and resampling techniques to address class imbalance. Historical data from cryptocurrency exchanges and data aggregators is collected, preprocessed, and used to train and evaluate these models. The impact of class imbalance, resampling techniques, and hyperparameter tuning on model performance is investigated. By analyzing historical cryptocurrency data, the methodology emphasizes hyperparameter tuning and backtesting, ensuring realistic model assessment. Results highlight the importance of addressing class imbalance and identify consistently outperforming models such as random forest, XGBoost, and gradient boosting. Our findings demonstrate that these models outperform others, indicating promising avenues for future research, particularly in sentiment analysis, reinforcement learning, and deep learning. This study provides valuable guidance for navigating the complex landscape of algorithmic trading in cryptocurrencies. By leveraging the findings and recommendations presented, practitioners can develop more robust and profitable trading strategies tailored to the unique characteristics of this emerging market.
Non-fungible tokens (NFTs) have gained mainstream attention in the fintech community, but there is little research on their statistical properties. This study investigates the long-memory characteristics of NFT returns and volatility, focusing on their potential for predicting price movements. As NFTs do not conform to traditional models, understanding their unique features is crucial for comprehending complex market dynamics. This study aims to reveal the impact of macroeconomic factors on NFT prices, understand their correlation and develop predictive models using autoregression and artificial intelligence (AI) technology. This research utilized datasets from the Centers for Disease Control and Prevention (CDC), U.S. Bureau of Labor Statistics, Bureau of Economic Analysis, Christieâs, Dune, and Google Trends. Correlation and p value tests revealed strong relationships between NFT prices and variables such as weekly volume, pandemics, inflation and security. The Baseline Model using autoregression with NFT volume, security and technology factors outperformed all other models demonstrating the speculative volatility of NFTs. The Transformer Model using transformers, an architecture used by ChatGPT, Gemini and Stable Diffusion, showed high accuracy with less feature selection and preprocessing efforts. This study provides a novelty using a systematic approach for researchers to perform financial forecasting and contributes to the scarce literature on NFTs. This research offers valuable insights to investors and private agents regarding the right economic conditions for NFT investments by reducing portfolio risks and making informed decisions. To the authorsâ best knowledge, this is the first study to utilize time-series transformers for forecasting NFTs based on macroeconomic factors.
Decentralization is a core principle of blockchain technology and Decentralized Autonomous Organizations (DAOs), enhancing security and resilience by distributing control across a network. Traditional metrics like the Gini coefficient and Nakamoto coefficient often fall short in capturing the complex dynamics of decentralization. This paper introduces the Apokedro decentralization index, a metric that evaluates decentralization by considering the probabilities of all possible subsets of nodes that could collectively centralize control. These concepts from game theory, such as the Nash equilibrium, and the Apokedro index, when incorporated, provide a nuanced assessment of centralization risks. Key contributions include the mathematical formulation of the index, an efficient computational algorithm utilizing pruning techniques, and benchmarking experiments that compare the index performance against traditional metrics across various statistical distributions. The Apokedro index offers a comprehensive tool for measuring decentralization in blockchain networks and DAOs.
Ahmed Bouteska, Taimur Sharif, Layal Isskandarani, Mohammad Zoynul Abedin
This research investigates how market-wide conditions (macro aspects) and individual cryptocurrency-specific characteristics (micro aspects) influence the efficiency of cryptocurrency markets. Macro aspects encompass the impacts of overall market liquidity, volatility, and global uncertainty events (e.g., the COVID-19 pandemic and geopolitical conflicts) on market efficiency. Micro aspects focus on cryptocurrency-specific attributes, such as liquidity and volatility levels, and their effects on price delays. Our findings reveal that rising liquidity and declining volatility enhance market efficiency at both macro and micro levels. Furthermore, we observe that during the periods of uncertainty, inefficiencies are exacerbated among less liquid and more volatile cryptocurrencies. We propose that the perceived uncertainties and substantial transaction costs associated with cryptocurrencies that lack liquidity and exhibit high volatility act as deterrents, diminishing the eagerness of active traders to participate in arbitrage trading. Consequently, this leads to inefficiencies in the market. The results of this study offer valuable insights for financial market regulators and authorities as well as investors associated with the crypto market, particularly during the times of financial turmoils.
The emergence of decentralized finance (DeFi) allows arbitrageurs to obtain risk-free income from price gaps of cryptocurrency tokens in many global markets. Several automated arbitrage techniques have been invented to profit from single or multiple platforms, including Centralized and Decentralized Exchange (CEX and DEX), triangular, and DEX-Fait. This paper proposes the arbitrage strategy of cross-cryptocurrency exchanges (ASCEX), a novel automated arbitrage strategy for CEX-DEX platforms, to maximize profit and loss (PNL) using a token route searching algorithm. Based on feature comparison, ASCEX outperforms the existing trading strategies available. Our actual trade experiment shows that ASCEX can generate up to 0.95% monthly risk-free profit compared to 0.34% trading on DEX alone.
Purpose This paper examines the relationship between the degree of information asymmetry among investors and the occurrence of bubbles in cryptocurrency markets. Design/methodology/approach The study applies the Philipps, Shi and Yu (PSY) methodology to identify bubbles in 74 cryptocurrencies from July 2014 to April 2021. Findings The findings indicate that there is a negative relationship between the degree of information asymmetry among investors and the number and duration of bubbles across cryptocurrencies. Originality/value This finding supports the riding-bubble argument of Asako et al. (2020), which suggests that when the information asymmetry among investors is high, rational investors are less certain about what irrational, inexperienced investors might decide. This strategic uncertainty leads rational investors to close out their positions more quickly, resulting in a shorter duration of the bubble and a reduced propensity for new bubbles to emerge. The studyâs findings hold regardless of the proxies used to measure information asymmetry and noise trading, cryptocurrency characteristics and regression model specifications.
Has the mean-variance framework become obsolete? In this paper, we replace traditional varianceâcovariance methods of portfolio optimisation with relative Tsallis entropy and mutual information measures. Its goal is to enhance risk management and diversification in complicated finance ecosystems. We utilize the S&P 500 and Bitwise 10 cryptocurrency indicesâ daily returns (2019â2024 data) and conduct our analysis to the year 2020 under extreme shocks. Many models were trained with different configurations, like mean-variance (MV), mean-entropy (ME), and mean-mutual information (MI) traders and their corresponding variants, using Sharpeâs ratio, Jensenâs alpha, and entropy value of risk (EVAR). The findings indicate that entropic models outperform conventional models in terms of diversification and, especially, extreme risk management. Because the appropriate normalization conditions often fail to be satisfied, we can informally see that after a recalibration of the effective frontier, we obtain from EVAR an accumulated resilience aspect to these rare events while also observing the great potential of entropy-based models to replicate non-linear dependencies between assets. The results show that models combining entropy and mutual information optimise the gainâloss ratio (GLR), providing stable diversification and improved risk management, while maximising returns in complex and volatile market environments.
Ki Anisa Zahria Salsabila, Andry Alamsyah, Nora Amelda Rizal
Decentralized Finance (DeFi) has transformed financial systems by facilitating peer-to-peer transactions without intermediaries, supported by blockchain explorers like Etherscan, allowing transparency and analyzing transactions. However, understanding DeFi's transactional and network dynamics remains limited. Previous studies only focused narrowly on specific metrics and lacked comparative analysis across different types of subsectors. This study implements network analysis as part of graph analytics to examine three business models of DeFi subsectorsâliquid staking represented by Lido (LDO), lending represented by Aave (AAVE), and Decentralized Exchange (DEX) represented by Uniswap (UNI)âto explore patterns of connectivity, wealth distribution, and market behavior. By analyzing over 1 million transaction records from August 2023 to July 2024, we reveal clustering behaviors, the influence of high wealth nodes, and network adaptability to market volatility using degree distribution, modularity, degree centrality, temporal density, and wealth distribution metrics. These insights provide valuable contributions to understanding DeFi network dynamics and offer practical implications for enhancing scalability and stability in a decentralized ecosystem.
This research employs Multifractal Detrended Fluctuation Analysis (MFDFA) to investigate multifractal properties in financial variables, including Bitcoin prices and economic indicators. Spanning 2019â2022, the analysis reveals multifractal scaling not only in Bitcoin prices, but also in economic indicators such as inflation rates and energy commodity prices. The non-linear singularity spectra unveil the multifaceted nature of scaling properties. Temporal analysis exposes intriguing trends in multifractality with implications for market efficiency. Furthermore, correlation analysis unveils connections among multifractal properties. For instance, a positive correlation between oil prices and Bitcoin suggests similar market forces. The log-log plot of fluctuation function Fq versus lag size demonstrates a power-law relationship, characteristic of multifractal systems. The empirical dataâs alignment in log-log space suggests self-similarity in the Bitcoin time series, supporting multifractality. The calculated Hurst exponents values suggest varying degrees of multifractality across the years, with 2021 exhibiting the highest degree and 2022 the lowest. Furthermore, an asymmetry index (0.5767) deviating from 0.5 indicates that the multifractal nature of the Bitcoin market is not symmetric. This research enhances risk assessment and portfolio optimization in finance. It challenges the Efficient Market Hypothesis (EMH), emphasizing the significance of MFDFA in comprehending financial market and economic factorâs relationships.
The aim of this research is to investigate the long-term relationships among the dollar exchange rate (TRY/USD), gold (GAU/USD), the Borsa Istanbul 100 Index (BIST 100) and the prices of Bitcoin (BTC/USD), Ethereum (ETH/USD), and Binance Coin (BNB/USD). Since the series contain structural breaks, Fourier unit root tests were used to model the structural breaks. As the method of this study, the relationships between variables in the long term were examined by using Fourier Shin (FSHIN) and Shin (1994) (SHIN) cointegration tests. The findings of this study showed that cryptocurrencies are cointegrated among themselves under structural breaks in the long term; investment instruments are cointegrated among themselves. In addition, as a result of this study, it was determined financial instruments and cryptocurrencies do not move in along over time under structural breaks.
This study aims to develop a dynamic portfolio trading system for high-risk profiles of cryptocurrencies in two phases: 1) portfolio selection and 2) portfolio construction. In the first phase, we propose a novel algorithmic trading model applying a Convolutional Neural Network (CNN) using a 2-D convolution layer with eight kernels of 3Ă3 sizes based on the prediction of selected technical indicators to predict buy/sell trading signals. To effectively increase the accuracy of the CNN model, first, the H-step ahead predictions of the selected technical indicators based on Long-short-term-memory (LSTM) along with the indicators themselves have been used to construct input matrices of the CNN model. A new price labeling approach was proposed to determine buying or selling points using the zigzag indicator (ZZ) in our CNN model. Assets with buy signals have been selected to construct the proposed portfolio. In the second phase, we propose a novel robust approach based on Holt-Winters-Multiplicative (HWM) to determine the realized crypto portfolio weights robustly by considering the seasonal effects. The experimental results show that our developed system outperforms the competing models for 30 cryptocurrencies with a high-risk profile in the two phases.
Mohammad Abdullah, Mohammad Ashraful Ferdous Chowdhury, G. M. Wali Ullah
This study inspects the asymmetric tail risk dynamics, efficiency, and interconnectedness among FinTech stocks, cryptocurrencies, and traditional assets. Firstly, we employ the Multifractal-Asymmetric Detrended Cross-Correlation Analysis to examine the cross-correlation patterns and efficiency dynamics of the analyzed assets. The findings reveal asymmetries in cross-correlations and the presence of multifractality, highlighting the nonlinear relationships among these assets and find FinTech assets are the most efficient. Secondly, we utilize the time domain quantile connectedness method to investigate tail risk connectedness, offering insights into the network's shock transmission and spillover effects. Our analysis identifies the major risk transmitters (FinTech stocks) and receivers (bond), emphasizing the interconnectedness of the assets. Additionally, the study conducts bivariate portfolio analysis, considering short and long investment horizons, to guide asset allocation and hedging strategies. Our findings have significant implications for facilitating informed investment strategies and improving the stability and resilience of financial markets.
Abstract Market efficiency assumes that prices in financial markets are perfectly informative and, therefore, it is not possible to design trading strategies that outperform the market. The concept of efficiency has important implications for financial stability and, consequently, for financial policies. If asset returns exhibit persistent or anti-persistent behavior, then predictability based on past returns might be possible, which would be a clear violation of the weak form of efficiency. Many studies rely on the Hurst exponent to evaluate the level of memory of financial returns, and the purpose of this paper is to show that long memory or anti-persistence of financial returns is not incompatible with the random walk model or the efficient market hypothesis (EMH). The use of the Hurst exponent to demonstrate the inefficiency of financial markets using common estimators is troublesome, especially when applied to financial returns, since values of $$\hat{H} \ne 0.5$$ <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"> <mml:mrow> <mml:mover> <mml:mi>H</mml:mi> <mml:mo>^</mml:mo> </mml:mover> <mml:mo>â </mml:mo> <mml:mn>0.5</mml:mn> </mml:mrow> </mml:math> are not evidence against the random walk model or the EMH. Moreover, the high variability of Hurst exponent estimates and their dependence on the chosen algorithm should motivate careful use of this tool. This study proposes a simple theoretical explanation and an extensive simulation study to show that $$\hat{H} \ne 0.5$$ <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"> <mml:mrow> <mml:mover> <mml:mi>H</mml:mi> <mml:mo>^</mml:mo> </mml:mover> <mml:mo>â </mml:mo> <mml:mn>0.5</mml:mn> </mml:mrow> </mml:math> for financial returns is perfectly compatible with the random walk model. As a robustness check, both the traditional rescaled range and the wavelet lifting algorithms are used. Applications to real data are also discussed to show that the empirical values of the Hurst exponent are in the range suggested by the simulations, providing evidence that over-reliance on the Hurst exponent could lead to erroneous rejection of the random walk model. Specifically, the paper presents an application to the daily returns of stock market indices (DJIA and S&P 500) over a period of more than 30 years and cryptocurrencies (Bitcoin and Ethereum) over a period of more than 5 years.
The purpose of this study is to conduct an empirical comparative study of volatility models for three of the most popular cryptocurrencies. We study the volatility of the following cryptocurrencies: Bitcoin, Ethereum, and Litecoin. We consider the GARCH-type, boosting-family-tree-based ensemble learning, and ANFIS volatility models for these financial crypto-assets, which some have claimed capture stylized facts about cryptocurrency volatility well. We conduct comparative studies on in-sample and out-of-sample empirical analyses. The results show that tree-based ensemble learning delivers better forecast accuracy. Nevertheless, the performance of some GARCH-type volatility models is relatively close to that of the best model on both training and evaluation samples.