In this study, we examined the regime-dependent dynamics and interrelationships among major cryptocurrencies, Bitcoin (BTC), Ethereum (ETH), and Monero (XMR), using high-frequency one-minute data from January 2020 to April 2025. To capture the presence of latent structural shifts without assuming Markovian transitions, we employed a Gaussian Mixture Model (GMM), which flexibly clustered distributions into two, empirically distinct regimes. Regime-specific Vector Autoregressive (VAR) models were then estimated to analyze interdependencies, spillovers, and shock transmission mechanisms across these digital assets. In the calm regime, the return dynamics were primarily self-driven, with limited cross-asset responses. Conversely, the volatile regime exhibited stronger and more persistent interlinkages, with BTC consistently acting as the principal transmitter of shocks to ETH and XMR, while ETH acts as a secondary transmitter, whereas XMR remains largely a risk recipient, absorbing external shocks with limited feedback into the system. These findings were corroborated through impulse response functions and forecast error variance decompositions, which consistently revealed asymmetric interdependence structures across the regimes. The Granger causality indicated more stable and statistically significant causal relationships in the calm regime than in the volatile regime. Furthermore, the Bai-Perron structural break tests confirmed the absence of significant deterministic breaks in the return series, reinforcing the validity of the GMM-based regime identification. These findings have practical implications for investors, regulators, and risk managers when modeling contagion and developing risk management strategies in cryptocurrency markets, especially during periods of heightened volatility.
Citation: Prashant Joshi. Regime-Specific interdependencies in cryptocurrency markets: A high-frequency GMM-VAR approach[J]. Data Science in Finance and Economics, 2025, 5(3): 419-439. doi: 10.3934/DSFE.2025017
In this study, we examined the regime-dependent dynamics and interrelationships among major cryptocurrencies, Bitcoin (BTC), Ethereum (ETH), and Monero (XMR), using high-frequency one-minute data from January 2020 to April 2025. To capture the presence of latent structural shifts without assuming Markovian transitions, we employed a Gaussian Mixture Model (GMM), which flexibly clustered distributions into two, empirically distinct regimes. Regime-specific Vector Autoregressive (VAR) models were then estimated to analyze interdependencies, spillovers, and shock transmission mechanisms across these digital assets. In the calm regime, the return dynamics were primarily self-driven, with limited cross-asset responses. Conversely, the volatile regime exhibited stronger and more persistent interlinkages, with BTC consistently acting as the principal transmitter of shocks to ETH and XMR, while ETH acts as a secondary transmitter, whereas XMR remains largely a risk recipient, absorbing external shocks with limited feedback into the system. These findings were corroborated through impulse response functions and forecast error variance decompositions, which consistently revealed asymmetric interdependence structures across the regimes. The Granger causality indicated more stable and statistically significant causal relationships in the calm regime than in the volatile regime. Furthermore, the Bai-Perron structural break tests confirmed the absence of significant deterministic breaks in the return series, reinforcing the validity of the GMM-based regime identification. These findings have practical implications for investors, regulators, and risk managers when modeling contagion and developing risk management strategies in cryptocurrency markets, especially during periods of heightened volatility.
| [1] |
Aarts E, Haslbeck JMB (2024) Modelling Psychological Time Series with Multilevel Hidden Markov Models: A Numerical Evaluation and Tutorial. PsyarXiv. https://doi.org/10.31234/osf.io/y2u5s doi: 10.31234/osf.io/y2u5s
|
| [2] |
Agakishiev I, Härdle WK, Becker D, et al. (2025) Regime switching forecasting for cryptocurrencies. Digit Financ 7: 107–131. https://doi.org/10.1007/s42521-024-00123-2 doi: 10.1007/s42521-024-00123-2
|
| [3] | Alzahrani S, Daim TU (2019) Analysis of the Cryptocurrency Adoption Decision: Literature Review. 2019 Portland International Conference on Management of Engineering and Technology (PICMET), 1–11. https://doi.org/10.23919/PICMET.2019.8893819 |
| [4] |
Ante L (2021) Smart contracts on the blockchain – A bibliometric analysis and review. Telemat Inform 57: 101519. https://doi.org/10.1016/j.tele.2020.101519 doi: 10.1016/j.tele.2020.101519
|
| [5] |
Bai J, Perron P (1998) Estimating and Testing Linear Models with Multiple Structural Changes. Econometrica 66: 47. https://doi.org/10.2307/2998540 doi: 10.2307/2998540
|
| [6] |
Bai J, Perron P (2003) Computation and analysis of multiple structural change models. J Appl Econometrics 18: 1–22. https://doi.org/10.1002/jae.659 doi: 10.1002/jae.659
|
| [7] | Bishop CM (2006) Pattern Recognition and Machine Learning. Springer. |
| [8] |
Bouri E, Lucey B, Roubaud D (2020) Cryptocurrencies and the downside risk in equity investments. Financ Res Lett 33: 101211. https://doi.org/10.1016/j.frl.2019.06.009 doi: 10.1016/j.frl.2019.06.009
|
| [9] |
Corbet S, Lucey B, Yarovaya L (2018) Datestamping the Bitcoin and Ethereum bubbles. Financ Res Lett 26: 81–88. https://doi.org/10.1016/j.frl.2017.12.006 doi: 10.1016/j.frl.2017.12.006
|
| [10] | Council F (2021) The Main Roadblocks to Crypto Moving Mainstream. Available from: https://www.forbes.com/sites/forbesbusinesscouncil/2021/06/23/the-main-roadblocks-to-crypto-moving-mainstream/?sh=2e629de922b9. |
| [11] |
Dempster AP, Laird NM, Rubin DB (1977) Maximum Likelihood from Incomplete Data Via the EM Algorithm. J R Stat Soc B 39: 1–22. https://doi.org/10.1111/j.2517-6161.1977.tb01600.x doi: 10.1111/j.2517-6161.1977.tb01600.x
|
| [12] |
Dickey DA, Fuller WA (1979) Distribution of the Estimators for Autoregressive Time Series with a Unit Root. J Am Stat Assoc 74: 427. https://doi.org/10.2307/2286348 doi: 10.2307/2286348
|
| [13] |
Dickey DA, Fuller WA (1981) Likelihood Ratio Statistics for Autoregressive Time Series with a Unit Root. Econometrica 49: 1057. https://doi.org/10.2307/1912517 doi: 10.2307/1912517
|
| [14] | Dimri SC, Indu R, Negi HS, et al. (2024) Hidden Markov Model—Applications, Strengths, and Weaknesses. 2024 2nd International Conference on Device Intelligence, Computing and Communication Technologies (DICCT), 300–305. https://doi.org/10.1109/DICCT61038.2024.10532827 |
| [15] | Eirola E, Lendasse A (2013) Gaussian Mixture Models for Time Series Modelling, Forecasting, and Interpolation. In: A. Tucker, F. Höppner, A. Siebes, & S. Swift (Eds.), Advances in Intelligent Data Analysis XII, 162–173. Springer Berlin Heidelberg. https://doi.org/10.1007/978-3-642-41398-8_15 |
| [16] |
Engle RF (1982) Autoregressive Conditional Heteroscedasticity with Estimates of the Variance of United Kingdom Inflation. Econometrica 50: 987. https://doi.org/10.2307/1912773 doi: 10.2307/1912773
|
| [17] |
Engle RF, Manganelli S (2004) CAViaR: Conditional Autoregressive Value at Risk by Regression Quantiles. J Bus Econ Stat 22: 367–381. https://doi.org/10.1198/073500104000000370 doi: 10.1198/073500104000000370
|
| [18] | Esmael B, Arnaout A, Fruhwirth RK, et al. (2012) Improving time series classification using Hidden Markov Models. 2012 12th International Conference on Hybrid Intelligent Systems (HIS), 502–507. https://doi.org/10.1109/HIS.2012.6421385 |
| [19] |
Figá-Talamanca G, Patacca M (2019) Does market attention affect Bitcoin returns and volatility? Decis Econ Financ 42: 135–155. https://doi.org/10.1007/s10203-019-00258-7 doi: 10.1007/s10203-019-00258-7
|
| [20] |
Figà-Talamanca G, Focardi S, Patacca M (2021) Regime switches and commonalities of the cryptocurrencies asset class. North Am J Econ Financ 57: 101425. https://doi.org/10.1016/j.najef.2021.101425 doi: 10.1016/j.najef.2021.101425
|
| [21] |
Foroni B, Merlo L, Petrella L (2024) Expectile hidden Markov regression models for analyzing cryptocurrency returns. Stat Comput 34: 66. https://doi.org/10.1007/s11222-023-10377-2 doi: 10.1007/s11222-023-10377-2
|
| [22] | Gershman SJ, Blei DM (2011) A Tutorial on Bayesian Nonparametric Models (Version 2). arXiv. https://doi.org/10.48550/arXiv.1106.2697 |
| [23] |
Gonçalves S, Kilian L (2004) Bootstrapping autoregressions with conditional heteroskedasticity of unknown form. J Econometrics 123: 89–120. https://doi.org/10.1016/j.jeconom.2003.10.030 doi: 10.1016/j.jeconom.2003.10.030
|
| [24] |
Granger CWJ (1969) Investigating Causal Relations by Econometric Models and Cross-spectral Methods. Econometrica 37: 424. https://doi.org/10.2307/1912791 doi: 10.2307/1912791
|
| [25] |
Guesmi K, Saadi S, Abid I, et al. (2019) Portfolio diversification with virtual currency: Evidence from bitcoin. Int Rev Financ Anal 63: 431–437. https://doi.org/10.1016/j.irfa.2018.03.004 doi: 10.1016/j.irfa.2018.03.004
|
| [26] |
Hadan H, Zhang-Kennedy L, Nacke L, et al. (2024) Comprehending the Crypto-Curious: How Investors and Inexperienced Potential Investors Perceive and Practice Cryptocurrency Trading. Int J Human–Comput Interact 40: 5675–5696. https://doi.org/10.1080/10447318.2023.2239556 doi: 10.1080/10447318.2023.2239556
|
| [27] |
Hamilton JD (1989) A New Approach to the Economic Analysis of Nonstationary Time Series and the Business Cycle. Econometrica 57: 357. https://doi.org/10.2307/1912559 doi: 10.2307/1912559
|
| [28] | Hamilton JD (1994) Time Series Analysis. Princeton University Press. |
| [29] |
Ji Q, Bouri E, Gupta R, et al. (2018) Network causality structures among Bitcoin and other financial assets: A directed acyclic graph approach. Q Rev Econ Financ 70: 203–213. https://doi.org/10.1016/j.qref.2018.05.016 doi: 10.1016/j.qref.2018.05.016
|
| [30] | Johnson MJ, Willsky AS (2013) Bayesian Nonparametric Hidden Semi-Markov Models. J Mach Learn Res 14: 673–701. Available from: https://www.jmlr.org/papers/volume14/johnson13a/johnson13a.pdf. |
| [31] |
Kalliovirta L, Meitz M, Saikkonen P (2016) Gaussian mixture vector autoregression. J Econometrics 192: 485–498. https://doi.org/10.1016/j.jeconom.2016.02.012 doi: 10.1016/j.jeconom.2016.02.012
|
| [32] |
Katsiampa P, Corbet S, Lucey B (2019) Volatility spillover effects in leading cryptocurrencies: A BEKK-MGARCH analysis. Financ Res Lett 29: 68–74. https://doi.org/10.1016/j.frl.2019.03.009 doi: 10.1016/j.frl.2019.03.009
|
| [33] |
Kim K, Lee SYT, Assar S (2022) The dynamics of cryptocurrency market behavior: Sentiment analysis using Markov chains. Ind Manage Data Syst 122: 365–395. https://doi.org/10.1108/IMDS-04-2021-0232 doi: 10.1108/IMDS-04-2021-0232
|
| [34] |
Kochliaridis V, Papadopoulou A, Vlahavas I (2024) UNSURE - A machine learning approach to cryptocurrency trading. Appl Intell 54: 5688–5710. https://doi.org/10.1007/s10489-024-05407-z doi: 10.1007/s10489-024-05407-z
|
| [35] |
Koenker R, Bassett G (1978) Regression Quantiles. Econometrica 46: 33. https://doi.org/10.2307/1913643 doi: 10.2307/1913643
|
| [36] |
Lee S, Kim C (2023) Dirichlet process mixture models using matrix‐generalized half‐t distribution. Stat 12: e599. https://doi.org/10.1002/sta4.599 doi: 10.1002/sta4.599
|
| [37] |
Li B (2023) Hidden Markov Model Based Stock Price Prediction: A Financial Research Report Based on Big Data Technology. SSRN Electronic J. https://doi.org/10.2139/ssrn.4622722 doi: 10.2139/ssrn.4622722
|
| [38] |
Ljung GM, Box GEP (1978) On a measure of lack of fit in time series models. Biometrika 65: 297–303. https://doi.org/10.1093/biomet/65.2.297 doi: 10.1093/biomet/65.2.297
|
| [39] |
Ma F, Liang C, Ma Y, et al. (2020) Cryptocurrency volatility forecasting: A Markov regime‐switching MIDAS approach. J Forecast 39: 1277–1290. https://doi.org/10.1002/for.2691 doi: 10.1002/for.2691
|
| [40] |
Merlo L, Petrella L, Raponi V (2021) Forecasting VaR and ES using a joint quantile regression and implications in portfolio allocation. J Bank Financ 133: 106248. https://doi.org/10.48550/arXiv.2106.06518 doi: 10.48550/arXiv.2106.06518
|
| [41] | Meunier S (2018) Blockchain 101. In: Transforming Climate Finance and Green Investment with Blockchains, 23–34. Elsevier. https://doi.org/10.1016/B978-0-12-814447-3.00003-3 |
| [42] |
Mgadmi N, Béjaoui A, Moussa W (2023) Disentangling the Nonlinearity Effect in Cryptocurrency Markets During the Covid-19 Pandemic: Evidence from a Regime-Switching Approach. Asia-Pac Financ Mark 30: 457–473. https://doi.org/10.1007/s10690-022-09384-6 doi: 10.1007/s10690-022-09384-6
|
| [43] |
Mgadmi N, Sadraoui T, Abidi A (2024) Causality between stock indices and cryptocurrencies before and during the Russo–Ukrainian war. Int Rev Econ 71: 301–323. https://doi.org/10.1007/s12232-023-00444-5 doi: 10.1007/s12232-023-00444-5
|
| [44] |
Morkūnaitė I, Celov D, Leipus R (2024) Evaluation of Value-at-Risk (VaR) using the Gaussian Mixture Models. Res Stat 2: 2346075. https://doi.org/10.1080/27684520.2024.2346075 doi: 10.1080/27684520.2024.2346075
|
| [45] |
Mungo L, Bartolucci S, Alessandretti L (2024) Cryptocurrency co-investment network: Token returns reflect investment patterns. EPJ Data Sci 13: 11. https://doi.org/10.1140/epjds/s13688-023-00446-x doi: 10.1140/epjds/s13688-023-00446-x
|
| [46] |
Oelschläger L, Adam T (2023) Detecting bearish and bullish markets in financial time series using hierarchical hidden Markov models. Stat Model 23: 107–126. https://doi.org/10.1177/1471082X211034048 doi: 10.1177/1471082X211034048
|
| [47] |
Pennoni F, Bartolucci F, Forte G, et al. (2022) Exploring the dependencies among main cryptocurrency log‐returns: A hidden Markov model. Economic Notes 51: e12193. https://doi.org/10.1111/ecno.12193 doi: 10.1111/ecno.12193
|
| [48] | Reynolds D (2009) Gaussian Mixture Models. In: S. Z. Li & A. Jain (Eds.), Encyclopedia of Biometrics, 659–663. Springer US. https://doi.org/10.1007/978-0-387-73003-5_196 |
| [49] |
Scrucca L (2024) Entropy-Based Volatility Analysis of Financial Log-Returns Using Gaussian Mixture Models. Entropy 26: 907. https://doi.org/10.3390/e26110907 doi: 10.3390/e26110907
|
| [50] |
Sims CA (1980) Macroeconomics and Reality. Econometrica 48: 1. https://doi.org/10.2307/1912017 doi: 10.2307/1912017
|
| [51] | Statista (2025) Biggest cryptocurrency in the world—both coins and tokens—based on market capitalization on February 25, 2025 (in billion U.S. dollars), Statista. Available from: https://www.statista.com/statistics/1269013/biggest-crypto-per-category-worldwide/. |
| [52] | Tan K, Chu M (2012) Estimation of Portfolio Return and Value at Risk Using a Class of Gaussian Mixture Distributions. Int J Bus Financ Res 6: 97–107. Available from: https://access.portico.org/stable?au=phx4jw4b1n4. |
| [53] |
Taylor JW (2019) Forecasting Value at Risk and Expected Shortfall Using a Semiparametric Approach Based on the Asymmetric Laplace Distribution. J Bus Econ Stat 37: 121–133. https://doi.org/10.1080/07350015.2017.1281815 doi: 10.1080/07350015.2017.1281815
|
| [54] |
Tengelin K, Sopasakis A (2020) Tick based clustering methodologies establishing support and resistance levels in the currency exchange market. Natl Account Rev 2: 354–366. https://doi.org/10.3934/NAR.2020021 doi: 10.3934/NAR.2020021
|
| [55] |
Turatti DE, Mendes FHPS, Mazzeu JHG (2025) Combining Volatility Forecasts of Duration‐Dependent Markov‐Switching Models. J Forecast 44: 1195–1210. https://doi.org/10.1002/for.3212 doi: 10.1002/for.3212
|
| [56] |
Wang Y, Andreeva G, Martin-Barragan B (2023) Machine learning approaches to forecasting cryptocurrency volatility: Considering internal and external determinants. Int Rev Financ Anal 90: 102914. https://doi.org/10.1016/j.irfa.2023.102914 doi: 10.1016/j.irfa.2023.102914
|
| [57] | Wang Y, Xu J, Huang SL, et al. (2025) Assessing Uncertainty in Stock Returns: A Gaussian Mixture Distribution-Based Method (Version 1). arXiv. https://doi.org/10.48550/arXiv.2503.06929 |
| [58] |
White H, Kim TH, Manganelli S (2015) VAR for VaR: Measuring tail dependence using multivariate regression quantiles. J Econometrics 187: 169–188. https://doi.org/10.1016/j.jeconom.2015.02.004 doi: 10.1016/j.jeconom.2015.02.004
|
| [59] |
Wong ACS, Chan WS (2005) Mixture Gaussian Time Series Modeling of Long-Term Market Returns. North Am Actuar J 9: 83–94. https://doi.org/10.1080/10920277.2005.10596227 doi: 10.1080/10920277.2005.10596227
|
| [60] |
Yi S, Xu Z, Wang GJ (2018) Volatility connectedness in the cryptocurrency market: Is Bitcoin a dominant cryptocurrency? Int Rev Financ Anal 60: 98–114. https://doi.org/10.1016/j.irfa.2018.08.012 doi: 10.1016/j.irfa.2018.08.012
|
| [61] |
Yuksel SE, Bolton J, Gader P (2015) Multiple-Instance Hidden Markov Models with Applications to Landmine Detection. Ieee T Geosci Remote 53: 6766–6775. https://doi.org/10.1109/TGRS.2015.2447576 doi: 10.1109/TGRS.2015.2447576
|
| [62] |
Zhang MH, Cheng QS (2003) Gaussian mixture modelling to detect random walks in capital markets. Math Comput Model 38: 503–508. https://doi.org/10.1016/S0895-7177(03)90022-7 doi: 10.1016/S0895-7177(03)90022-7
|
| [63] |
Zou Y, Lin Y, Song X (2024) Bayesian Heterogeneous Hidden Markov Models with an Unknown Number of States. J Comput Graph Stat 33: 15–24. https://doi.org/10.1080/10618600.2023.2231055 doi: 10.1080/10618600.2023.2231055
|
 DSFE-05-03-017-s001.pdf |
 |
Figures(4) / Tables(6)
Prashant Joshi. Regime-Specific interdependencies in cryptocurrency markets: A high-frequency GMM-VAR approach[J]. Data Science in Finance and Economics, 2025, 5(3): 419-439. doi: 10.3934/DSFE.2025017
DownLoad: