Chappell 2018
“Regime heteroskedasticity in Bitcoin: A comparison of Markov switching models” (Daniel Chappell, MPRA Paper 90682, University Library of Munich, 2018) is among the first papers to apply Markov regime-switching / hidden Markov models to a cryptocurrency return series. It fits 2-to-7-state MRS estimations to Bitcoin returns and, judged by BIC, Hannan-Quinn and AIC, finds a restricted five-state model best captures the data, which exhibits volatility clustering, volatility jumps, asymmetric volatility transitions and shock persistence. It appears in this vault as an early instance of the Markov Regime-Switching Model applied to the Cryptocurrency Market in a purely descriptive mode — a goodness-of-fit / regime-characterisation study with no trading strategy, no backtest, no transaction costs and no P&L. Its profitability grade is therefore inconclusive: it confirms crypto’s regime structure is real and HMM-describable, but makes — and supports — no profitability claim.
Connections
- Cryptocurrency Market — detects_regime, 5-state model best fits Bitcoin volatility regimes, source: https://ideas.repec.org/p/pra/mprapa/90682.html
- Markov Regime-Switching Model — proposes_model, m-state MRS/HMM fitted to Bitcoin returns, source: https://ideas.repec.org/p/pra/mprapa/90682.html
- Hidden Markov Model Regime Detection — detects_regime, MRS models equated with HMMs for regime heteroskedasticity, source: https://ideas.repec.org/p/pra/mprapa/90682.html
- Regime Classification — detects_regime, descriptive regime characterisation with no trading rule, source: https://ideas.repec.org/p/pra/mprapa/90682.html
- State-Count Selection — relates, optimal state count chosen by BIC/HQ/AIC information criteria, source: https://ideas.repec.org/p/pra/mprapa/90682.html
Sources
- Chappell, D. (2018). “Regime heteroskedasticity in Bitcoin: A comparison of Markov switching models.” MPRA Paper 90682, University Library of Munich, Germany — https://ideas.repec.org/p/pra/mprapa/90682.html