Paper: arXiv 2507.00575
Authors: Milan Pontiggia
Abstract
We assess the applicability of rough volatility models to Bitcoin realized volatility using the normalised p-variation framework of Cont and Das (2024). Applying this model-free estimator to high-frequency Bitcoin data from 2017 to 2024 across multiple sampling resolutions, we find that the normalised statistic remains strictly negative, precluding the estimation of a valid roughness index. Stationarity tests and robustness checks reveal no significant evidence of non-stationarity or structural breaks as explanatory factors. Instead, convergent evidence from three complementary diagnostics, namely Multifractal Detrended Fluctuation Analysis, log-log moment scaling, and wavelet leaders, reveals a multifractal structure in Bitcoin volatility. This behaviour violates the homogeneity assumptions underlying rough volatility estimation and accounts for the estimator’s systematic failure. These findings suggest that while rough volatility models perform well in traditional markets, they are structurally misaligned with the empirical features of Bitcoin volatility.
Complexity vs Empirical Score
- Math Complexity: 8.5/10
- Empirical Rigor: 9.0/10
- Quadrant: Holy Grail — high math complexity, high empirical rigor
Why this score: The paper employs advanced mathematical frameworks like the normalised p-variation estimator, multifractal analysis, and scaling laws, indicating high math complexity. It also demonstrates strong empirical rigor by using extensive high-frequency data (2017-2024), conducting robustness checks, multiple diagnostics (DFA, log-log scaling, wavelets), and addressing model assumptions and data limitations.
Research Flowchart
flowchart TD
A["Research Goal: Assess applicability of rough volatility models to Bitcoin using Cont and Das (2024) normalised p-variation estimator"] --> B["Data Input: High-frequency Bitcoin data 2017-2024"]
B --> C["Core Methodology: Apply normalised p-variation across multiple sampling resolutions"]
C --> D{"Is the normalised statistic positive?"}
D -- No --> E["Estimation Failure: Statistic strictly negative precludes roughness index"]
D -- Yes --> F["Roughness Index Estimated"]
E --> G{"Check Explanatory Factors"}
G -- Non-stationarity/Structural Breaks --> H["Disconfirmed: Stationarity tests show no significant evidence"]
G -- Multifractal Structure --> I["Confirmed: Convergent evidence from MF-DFA, log-log moment scaling, and wavelet leaders"]
H --> I
I --> J["Key Finding: Bitcoin volatility exhibits multifractality, violating homogeneity assumptions required for rough volatility models"]
J --> K["Implication: Rough volatility models are structurally misaligned with Bitcoin volatility features"]
F --> K
Evidence Card
What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card
Proposes
The paper applies the model‑free normalised p‑variation estimator of Cont and Das (2024) to Bitcoin realised volatility and complements it with multifractal diagnostics.
Source sentence
We assess the applicability of rough volatility models to Bitcoin realized volatility using the normalised p-variation framework of Cont and Das (2024).
Assumes 5
- The normalised p‑variation method requires homogeneous scaling of the volatility process. model
Source sentence
This behaviour violates the homogeneity assumptions underlying rough volatility estimation and accounts for the estimator's systematic failure.
- Bitcoin realised volatility is stationary and free of structural breaks. statistical
Source sentence
Stationarity tests and robustness checks reveal no significant evidence of non-stationarity or structural breaks as explanatory factors.
- The high‑frequency minute‑level price data from Bitstamp are of sufficient quality for volatility estimation over 2017‑2024. data
Source sentence
To mitigate these issues, we restrict our analysis to the 2017-2024 period, during which data quality is sufficiently stable for high-frequency volatility estimation.
- The dataset consists of BTC/USD prices recorded at one‑minute intervals. data
Source sentence
3.2 Data Description and Preprocessing This study employs a high-frequency dataset of BTC/USD prices recorded at one-minute intervals, derived from the publicly available “Bitcoin Historical Data” dataset on Kaggle.
Show 1 more
- Absolute returns can be used as a proxy for instantaneous volatility for the normalised p‑variation estimator. model
Source sentence
Accordingly, we estimate the H¨older roughness of the volatility path itself (as captured by absolute returns), rather than of the price or log-price process.
Data
- Bitcoin Historical Data market-prices public
Source sentence
3.2 Data Description and Preprocessing This study employs a high-frequency dataset of BTC/USD prices recorded at one-minute intervals, derived from the publicly available “Bitcoin Historical Data” dataset on Kaggle. The Bitcoin high-frequency price data are publicly accessible from Kaggle at https://www.kaggle.com/datasets/mczielinski/bitcoin-historical-d ata.
Costs
Not applicable
Finds 5
- The normalised statistic log W(L,K,p) remains strictly negative across all configurations, precluding estimation of a valid roughness index. statistical
Source sentence
Applying this model-free estimator to high-frequency Bitcoin data from 2017 to 2024 across multiple sampling resolutions, we find that the normalised statistic remains strictly negative, precluding the estimation of a valid roughness index. The statistic log W(L, K, p) remains strictly negative across all configurations, precluding the estima-tion of a valid roughness index.
- Stationarity diagnostics and robustness checks find no significant evidence of non‑stationarity or structural breaks. statistical
Source sentence
Stationarity tests and robustness checks reveal no significant evidence of non-stationarity or structural breaks as explanatory factors. Diagnostic tests rule out non-stationarity, structural breaks, and sampling artefacts as plausible explanations.
- Multifractal Detrended Fluctuation Analysis, log‑log moment scaling, and wavelet leaders consistently indicate a robust multifractal temporal structure in Bitcoin realised volatility. statistical
Source sentence
The MF-DFA, log-log moment scaling, and wavelet leaders analyses consistently in-dicate that Bitcoin realized volatility exhibits a multifractal temporal structure. Instead, consistent evidence from MF-DFA, log-log moment scaling, and wavelet leaders indicates that Bitcoin realized volatility follows a robust multifractal temporal structure, driven by dynamic correlations rather than marginal distributional effects.
- The observed multifractality violates the homogeneity assumptions required by the normalised p‑variation framework, explaining the failure to identify a well‑defined scaling exponent. theoretical
Source sentence
This violates the ho-mogeneity assumptions required by the normalised p-variation framework and provides a structural explanation for the failure to identify a well-defined scaling exponent. This violates the homogeneity assumptions required by rough volatility models and accounts for the systematic failure of the estima-tor.
- Rough volatility models are structurally incompatible with Bitcoin and are likely to produce unreliable forecasts and pricing outcomes. negative
Source sentence
Consequently, rough volatility models are struc-turally incompatible with Bitcoin and are likely to produce unreliable forecasts and pric-ing outcomes.
Checked by others
Supported by
- What is mature and what is still emerging in the cryptocurrency market?
Compare the findings
This paper: Multifractal Detrended Fluctuation Analysis, log‑log moment scaling, and wavelet leaders consistently indicate a robust multifractal temporal structure in Bitcoin realised volatility.
That paper: Return distributions, volatility clustering, and temporal multifractal correlations of the largest‑cap cryptocurrencies largely follow those observed in well‑established financial markets.
- Chance or Chaos? Fractal geometry aimed to inspect the nature of Bitcoin
Compare the findings
This paper: Multifractal Detrended Fluctuation Analysis, log‑log moment scaling, and wavelet leaders consistently indicate a robust multifractal temporal structure in Bitcoin realised volatility.
That paper: Bitcoin exhibits a high degree of persistence and long memory, indicating self‑similar fractal behaviour across frequencies and time intervals.
Qualified by
- Multifractality and sample size influence on Bitcoin volatility patterns
Compare the findings
This paper: Multifractal Detrended Fluctuation Analysis, log‑log moment scaling, and wavelet leaders consistently indicate a robust multifractal temporal structure in Bitcoin realised volatility.
That paper: The strength of multifractality in the realized volatility is smaller than that of the price‑return series and varies over time.
Code
No repository by the authors found in the paper.
Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.
Cited References (21)
- Volatility is Rough (2018) — link
- Consistent Estimation for Fractional Stochastic Volatility Model under High-Frequency Asymptotics (2022) — link
- Rough Volatility: Fact or Artefact? (2024) — link
- Trading and Arbitrage in Cryptocurrency Markets (2020) — link
- Volatility and Return Jumps in Bitcoin (2018) — link
- Modelling Volatility of Cryptocurrencies Using Markov-Switching GARCH Models (2019) — link
- Chaos, Randomness and Multi-Fractality in Bitcoin Market (2018) — link
- Multifractal Detrended Fluctuation Analysis of Nonstationary Time Series (2002) — link
and 13 more
- Bootstrap for Empirical Multifractal Analysis (2007) — link
- Decoupling the Short- and Long-Term Behavior of Stochastic Volatility (2022) — link
- Power and Bipower Variation with Stochastic Volatility and Jumps (2004) — link
- Rough Volatility of Bitcoin (2020) — link
- Multifractality and Sample Size Influence on Bitcoin Volatility Patterns — → 2511.03314
- Surrogate Time Series (2000) — link
- Components of Multifractality in the Central England Temperature Anomaly Series (2013) — link
- Bitcoin Historical Data (2025) — link
- Tests for Hurst Effect (1987) — link
- Estimates of the Regression Coefficient Based on Kendall’s Tau (1968) — link
- Distribution of the Estimators for Autoregressive Time Series with a Unit Root (1979) — link
- Testing the Null Hypothesis of Stationarity Against the Alternative of a Unit Root: How Sure Are We That Economic Time Series Have a Unit Root? (1992) — link
- Selective Review of Offline Change Point Detection Methods (2020) — link
1 of 21 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.