Oana Panazan, Catalin GHEORGHE, Aamir Aijaz Syed, Ahmed Jeribi
This study examines the dynamic interactions between precious metals, cryptocurrencies, stablecoins, safe-haven currencies, and two key macroeconomic indicators, the 5-year breakeven inflation expectation (T5YIE) and the 10-year minus 3-month Treasury yield spread (T10Y3M), over January 2016–July 2025. To capture nonlinear and multi-scale dependencies, the study applies Quantile-on-Quantile Regression (QQR) in combination with wavelet coherence (WCO) and wavelet transform coherence (WTC). The results indicate that major cryptocurrencies such as Bitcoin and Ethereum do not display robust or systematic links with inflation expectations or recession risk, limiting their role as macro-financial hedges. By contrast, the Japanese yen and Swiss franc show pronounced tail sensitivities, reaffirming their safe-haven status, while gold and its tokenized counterparts (DGX, PAXG) exhibit persistent long-run coherence with inflation expectations. Stablecoins demonstrate unstable short-term linkages shaped by liquidity shocks and market frictions. The research provides new evidence on the heterogeneous roles of digital and traditional assets in shaping macroeconomic expectations. The findings carry implications for investors, who should continue to rely on gold and safe-haven currencies for crisis hedging, and for regulators concerned with the systemic stability of emerging digital instruments.
We document significant reversal in cryptocurrency returns at 8-and 10-week horizons, concentrated in midsize, relatively volatile assets. Using a panel of 70 USDT-quoted tokens on Binance from January 2021 through March 2026, we show that a contrarian strategy of buying past losers and selling past winners, by forming Jegadeesh-Titman (1993) calendar-time overlapping portfolios, earns a 39.6% annualized return (Sharpe 0.96, Newey-West t = 2.10). Reversal is stronger among high-volatility assets, generating a Sharpe ratio of 1.37 (t = 3.19), and is strengthened outside of the largest assets, generating a Sharpe ratio of 1.69 (t = 3.80). The effect is robust across tercile, quintile, and decile sorts; skip period variants; inverse-volatility weighting; and temporal subsamples. A circular block bootstrap with 10,000 replications corroborates the high-volatility result nonparametrically with 95% of Sharpe ratios above 0.67, and the high-versus-low volatility gap positive in 94% of replications. Several economic mechanisms to rationalize these findings are discussed, including the tendency of treasury managers to sell into upswings, and Nagel's (2012) theory that reversal compensates liquidity providers.
This study, among many others, provides an initial quantitative contribution to the emerging literature as well as existing empirical evidence regarding contagion risks across cryptocurrency markets over time. Using VAR (Vector Autoregressive) and SVAR (Structural Vector Autoregressive) models with Granger causality, along with Student’s t-copulas, we find that Bitcoin is likely to act as an independent asset in this market, while Ripple and Litecoin tend to be recipients of contagion effects, and Ethereum appears to be a primary source of contagion. Our study offers additional insight into the investigation of contagion risks between both historical and future cryptocurrency values by employing Student’s t-copulas for joint distribution analysis and aims to determine whether these contagion effects remain consistent over time. The results suggest, in both cases, that all cryptocurrencies tend to move negatively in extreme value conditions. Investors are therefore encouraged to pay closer attention to “bad news” and market movement patterns in order to make timely decisions regarding buying, holding, and selling. Note: This thesis reflects the state of cryptocurrency markets and related quantitative models as of 2019-2021. The findings and conclusions are intended to preserve the integrity of the research conducted during this specific time period. I acknowledges that this field evolves rapidly, and an updated analysis may be presented in a forthcoming research paper.
This thesis examines whether cryptocurrencies can function as diversification or risk-reducing assets relative to the Swedish equity market during periods of financial stress. Using daily data for Bitcoin, Ethereum and Ripple from 2018 to 2024, their dynamic relationship with the OMX30 index is analyzed. To provide a broader benchmark, gold, the German DAX index, and the U.S. S&P 500 index are included as comparison assets. Periods of financial stress are identified as episodes in which the OMX30 declines by at least 10 percent from a recent peak. Time-varying correlations are estimated using a Dynamic Conditional Correlation GARCH (DCC-GARCH) model, allowing the analysis of how interasset relationships evolve over time. In addition, hedge effectiveness measures are employed to assess the cryptocurrencies practical ability to reduce portfolio risk.The results show that Bitcoin, Ethereum and Ripple exhibit weak but positive correlations with the Swedish equity market, implying that they may serve as diversifiers but not ashedges or safe-havens. During periods of financial stress, correlations tend to increase rather than decrease, indicating limited protective properties. Hedge effectiveness estimates further suggest that the risk-reducing capacity of cryptocurrencies is unstable and generally weak. Incontrast, gold displays more consistent negative correlations and superior hedging performance. Overall, the findings suggest that cryptocurrencies offer limited diversification benefits for Swedish investors and should not be considered reliable risk-mitigating assets during market stress.
This study investigates time-frequency volatility transmission of both cross-exchange instrument networks and cross-instrument exchange networks of crypto-markets under a unified time-varying parameter vector autoregression (TVP-VAR) framework. 5-minute closing prices of 16 variables throughout the 2021-2025 period are sampled to form the two types of networks by instrument and by exchange, comprising spot and perpetual instruments of Bitcoin (BTC) and Ethereum (ETH) on Binance, OKX, Bitget, and Bitfinex. The hourly logarithmic realized volatility is aggregated to estimate both the time- and frequency-domain connectedness partitioned into short- (1-8 hours), mid- (8-24 hours), and long-term (beyond 24 hours) components. The results indicate that the total connectedness of instrument networks uniformly exceeds that of exchange networks. The long-term component prevails, while the short-term component remains non-negligible with episodic amplification. Bitfinex functions as the sole net receiver across instrument networks, whereas Binance, OKX, and Bitget act as net transmitters. Within exchange networks, ETH instruments more frequently transmit volatility to BTC ones beyond eight hours, while this direction reverses within the 8-hour horizon. Spot instruments dominate perpetuals on OKX, Bitget, and Bitfinex with Binance as the main exception. Eight historical events elicit heterogeneous response modes contingent on the nature of the shocks, with the LUNA collapse and FTX crisis as the most influential in-sample events. The findings carry implications for cross-venue risk monitoring and horizon-aware portfolio management.
This paper studies volatility prediction for Ethereum in the post-Merge era. Using daily ETH/USD returns from 15 September 2022 to 23 April 2026, we compare standard GARCH(1,1), Heston-Nandi GARCH(1,1), cross-validated and aggregated EWMA predictors, and Nadaraya-Watson kernelregression predictors. The kernel forecasts are constructed from a rank-transformed state vector that captures recent volatility and signed-return conditions, allowing the conditional variance function to be nonlinear and state dependent. The results show that forecast performance is strongly horizon dependent. At the one-day horizon, the kernel predictor using the fitted GARCH volatility state delivers the lowest final cumulative squared prediction error, outperforming the standard GARCH benchmark and all EWMA-type competitors. At the ten-day-ahead horizon, the advantage of local nonparametric information weakens, and the mean-reverting structure of GARCH becomes more valuable. The estimated kernel surface reveals that predicted ETH volatility is highest when elevated recent volatility coincides with negative signed-return pressure. Conditional quantile results further show that kernel-based VaR improves lower-tail risk forecasts, especially at the 1% quantile. Overall, the evidence suggests that post-Merge Ethereum volatility is persistent, asymmetric, heavytailed, and nonlinear, and is best modelled by combining economically meaningful volatility states with flexible nonparametric forecasting maps.
Although many academic studies have examined volatility spillovers and dynamic correlations between stock markets, they have largely overlooked the perspective of Small and Medium-Sized Enterprise (SME) markets. On this basis, this study explores the interconnectedness and volatility correlation between Decentralized Finance (DeFi) markets and SME markets. To understand the correlation between these markets, we empirically analyse six European SME market indices—the FTSE AIM All-Share Index (AIM), BIST SME Industrial Index (BISTSME), Euronext Growth All-Share Index (EURONEXT), First North All-Share Index (FIRSTNORTH), IBEX Medium Cap Index (IBEXC), and Scale All-Share Performance Index (SCALE)—alongside three cryptocurrencies: Aave (AAVE), Ethereum (ETH), and Uniswap (UNI); two stablecoins: Dai (DAI) and USD Coin (USDC); and one synthetic asset: Synthetix (SNX). The study employed BEKK-GARCH and DCC-GARCH to analyse the existence of spillover effects and correlations from October 5, 2020, to August 18, 2024. The findings indicate that AAVE, ETH, and UNI, in particular, transmit significant volatility to the EURONEXT and FIRSTNORTH markets. However, bidirectional volatility spillover was detected between EURONEXT and AAVE, ETH, UNI, USDC, and SNX, and FIRSTNORTH and AAVE, ETH, UNI, and SNX. This suggests volatility interdependence between these markets and the existence of potential risk contagion channels.
Purpose We investigate the presence of contagion between Bitcoin and four traditional assets (stocks, bonds, gold and the US dollar exchange rate) over the period 2015–2024. Design/methodology/approach We implement a framework that combines the DCC-GARCH specification and a time-varying causal inference methodology. Findings Our findings support that Bitcoin remains weakly connected to the global financial markets. Contagion is limited, appearing sporadically from S&P 500 to Bitcoin and from Bitcoin to the US dollar index. However, when we impose a stricter definition of extreme correlation or a multivariate VAR specification, the contagion results vanish, indicating no systematic contagion between Bitcoin and traditional assets. Practical implications Our evidence implies that Bitcoin may be used as a useful portfolio diversification instrument. Originality/value We deploy a recently developed novel methodology that combines the DCC-GARCH model and a recent time-varying Granger causality procedure to distinguish between extreme high correlation and contagion and find no evidence of systematic contagion effects of Bitcoin with conventional asset classes.
This study investigates the relationship between dirty and clean cryptocurrencies and traditional stock index returns using the Quantile-Quantile (QQR) and Quantile-Quantile Granger Causality (QQGC) methods. The analyses were conducted using daily data from January 2018 to May 2025. QQR results show both positive and negative relationships between dirty and clean cryptocurrencies and the returns of the S&P 500, FTSE 100, TSX, and ASX indices at the low, medium, and high quantiles. According to the QQGC results, both dirty and clean cryptocurrencies showed predictive power for the returns of the S&P 500, FTSE 100, TSX, and ASX indices at different quantiles. Furthermore, it was found that both dirty and clean cryptocurrencies exhibit strong predictive power for S&P 500 and FTSE 100 returns, particularly in the middle quantiles. The results obtained reveal that distinguishing between dirty and clean cryptocurrencies under different market conditions provides important insights for investors' portfolio diversification strategies and risk management practices.
This study introduces a unified and methodologically symmetric comparative framework for multivariate cryptocurrency forecasting, addressing long-standing inconsistencies in prior research where model families, feature sets, and preprocessing pipelines differ across studies. Under an identical and rigorously controlled experimental setup, we benchmark six deep learning architectures—LSTM, GPT-2, Informer, Autoformer, Temporal Fusion Transformer (TFT), and a Vanilla Transformer—together with four widely used econometric models (ARIMA, VAR, GARCH, and a Random Walk baseline). All models are evaluated using a shared multivariate feature space composed of more than forty technical indicators, identical normalization procedures, harmonized sliding-window formations, and aligned temporal splits across five high-liquidity assets (BTC, ETH, XRP, XLM, and SOL). The experimental results show that transformer-based architectures consistently outperform both the recurrent baseline and classical econometric models across all assets. This superiority arises from the ability of attention mechanisms to capture long-range temporal dependencies and adaptively weight informative time steps, whereas recurrent models suffer from vanishing-gradient limitations and restricted effective memory. The best-performing deep learning models achieve MAPE values of 0.0289 (BTC, GPT-2), 0.0198 (ETH, Autoformer), 0.0418 (XRP, Informer), 0.0469 (XLM, Informer), and 0.0578 (SOL, TFT), substantially improving upon the performance of both LSTM and all econometric baselines. These findings highlight the effectiveness of attention-based architectures in modeling volatility-driven nonlinear dynamics and establish a reproducible, symmetry-preserving benchmark for future research in deep-learning-based financial forecasting.
Abstract This paper considers option valuation under finite mixture models in a discrete-time economy. Specifically, the Esscher transform is employed to select a pricing kernel. Novel finite mixture models with negative-shifted Gamma and negative-shifted inverse Gaussian distributions are developed. A hybrid finite mixture model that allows different parametric forms for component distributions is introduced to incorporate model uncertainty. An empirical characteristic function estimation method is employed to estimate the finite mixture models. Closed-form pricing formulas for a European call option are obtained for some finite mixture models. Empirical examples using data on the Bitcoin-USD prices are provided to illustrate an application of the proposed models to value Bitcoin options.
The aim of this dissertation is to test the applicability of two strategies – Dollar Cost Average (DCA) and Lump-Sum (LS) – in the context of the crypto market. We tested these strategies on three assets, namely Bitcoin, Ethereum and Ripple. We developed a simulation using daily historical data recorded over a period of nine years. We then calculated performance ratios and created an AR-GARCH model to analyse their properties and predictive capacity more effectively. Our empirical results show that all assets are highly volatile and exhibit heavy tails and asymmetry. Additionally, they are moderately to highly correlated with each other. We also presented proof of higher Sharpe and Sortino ratios for DCA strategies, with Bitcoin performing better than the other two assets. The results also show that Bitcoin has low-to-moderate shock sensitivity and high persistence; Ethereum has low shock sensitivity and high persistence; and Ripple has both high shock sensitivity and persistence. Furthermore, we observed the impact of strategy choice on volatility. When compared to DCA, LS lowered shock sensitivity in Bitcoin and Ripple, enhancing persistence, while having an insignificant effect on Ethereum. Finally, we demonstrate that our model exhibits superior predictive capacity with regard to Ripple compared to Bitcoin and Ethereum, and that all three assets are inefficient. These findings contribute to previous literature by providing novel empirical data and attesting to the attributes of cryptocurrencies. Furthermore, this thesis improves financial awareness and provides investors with valuable information.
Abstract This study explores the dynamic volatility spillovers and interconnectedness between cryptocurrency and traditional futures markets. Using a multi-method approach that integrates wavelet coherence analysis, TVP-VAR connectedness, and DCC-GARCH modeling, the research identifies notable shifts in spillover patterns during crises, such as the COVID-19 pandemic, the FTX collapse, and the Russia-Ukraine conflict. The results reveal that the correlations between Bitcoin futures and traditional asset classes depend on the market conditions and intensify during crises. The connectedness analysis shows that Bitcoin futures play a dual role, acting as a transmitter of long-term shocks and a receiver of short-term shocks during periods of crisis. Equity futures emerged as the primary long-term transmitters of shocks, whereas other assets acted as shock receivers during the pandemic. Furthermore, the study evaluates hedge ratios and portfolio weights using the DCC-GARCH model. The portfolio analysis reveals that Bitcoin futures require a minimal allocation within diversified portfolios, suggesting their limited effectiveness as a hedge and safe-haven asset. These results aim to inform portfolio managers in developing efficient hedging strategies and assist regulators in monitoring financial market stability. This study fills gaps in the existing literature by understanding how decentralized financial instruments interact with financial markets and providing insights into risk management in modern markets.
Abstract This study evaluates the predictive accuracy of traditional time series (TS) models versus machine learning (ML) methods in forecasting realized volatility across major cryptocurrencies—Bitcoin (BTC), Ethereum (ETH), Litecoin (LTC), and Ripple (XRP). Employing high-frequency data, we analyze cross-cryptocurrency volatility dynamics through two complementary approaches: volatility forecasting and connectedness analysis. Our findings reveal three key insights: (i) TS models, particularly the heterogeneous autoregressive (HAR) model, exhibit superior predictive performance over their ML counterparts, with the long short-term memory (LSTM) model providing competitive yet inconsistent results due to overfitting and short-term volatility challenges; (ii) including lagged realized volatility of large-cap coins improves predictive accuracy for mid-cap coins, especially XRP, whereas forecasts for large-cap coins remain stable, indicating more resilient volatility patterns; and (iii) volatility connectedness analysis reveals substantial spillover effects, particularly pronounced during market turmoil, with large-cap assets (BTC and ETH) acting as primary volatility transmitters and mid-cap assets (XRP and LTC) serving as volatility receivers. These results contribute to the understanding of volatility forecasting and risk management in cryptocurrency markets, offering implications for investors and policymakers in managing market risk and interdependencies in digital asset portfolios.
This study investigates the application of the Light Gradient Boosting Machine (LGBM) model for both deterministic and probabilistic forecasting of Bitcoin realized volatility. Utilizing a comprehensive set of 69 predictors -- encompassing market, behavioral, and macroeconomic indicators -- we evaluate the performance of LGBM-based models and compare them with both econometric and machine learning baselines. For probabilistic forecasting, we explore two quantile-based approaches: direct quantile regression using the pinball loss function, and a residual simulation method that transforms point forecasts into predictive distributions. To identify the main drivers of volatility, we employ gain-based and permutation feature importance techniques, consistently highlighting the significance of trading volume, lagged volatility measures, investor attention, and market capitalization. The results demonstrate that LGBM models effectively capture the nonlinear and high-variance characteristics of cryptocurrency markets while providing interpretable insights into the underlying volatility dynamics.
This research investigates the dynamics of connectedness among cryptocurrency and various risk factors, including oil price demand and supply shocks, EPU, GPR, and the ADS business conditions index using the quantile time-frequency connectedness approach. The findings reveal that cryptocurrency behaves as a net receiver of shocks in the short term but transitions to a net transmitter over the long term. Critical sources of both short- and long-term shocks are attributed to oil price demand, supply fluctuations, and GPR. However, during extreme events like the COVID−19 pandemic and the Russia-Ukraine war, cryptocurrency, oil shocks, and other indices alternately become net transmitters and receivers of shocks depending on time frames and quantile ranges. During periods of heightened market uncertainty, monitoring the interconnected behavior of these variables is critical for investors and policymakers aiming to predict market shifts and manage risks effectively.