Paper: arXiv 2508.07151

Authors: Roshan Shah

Abstract

We introduce a modular framework that extends the signature method to handle American option pricing under evolving volatility roughness. Building on the signature-pricing framework of Bayer et al. (2025), we add three practical innovations. First, we train a gradient-boosted ensemble to estimate the time-varying Hurst parameter H(t) from rolling windows of recent volatility data. Second, we feed these forecasts into a regime switch that chooses either a rough Bergomi or a calibrated Heston simulator, depending on the predicted roughness. Third, we accelerate signature-kernel evaluations with Random Fourier Features (RFF), cutting computational cost while preserving accuracy. Empirical tests on S&P 500 equity-index options reveal that the assumption of persistent roughness is frequently violated, particularly during stable market regimes when H(t) approaches or exceeds 0.5. The proposed hybrid framework provides a flexible structure that adapts to changing volatility roughness, improving performance over fixed-roughness baselines and reducing duality gaps in some regimes. By integrating a dynamic Hurst parameter estimation pipeline with efficient kernel approximations, we propose to enable tractable, real-time pricing of American options in dynamic volatility environments.

Complexity vs Empirical Score

  • Math Complexity: 8.0/10
  • Empirical Rigor: 5.5/10
  • Quadrant: Holy Grail — high math complexity, high empirical rigor

Why this score: The paper employs advanced mathematical concepts like rough-path theory, signature methods, and fractional Brownian motion with detailed formulas and derivations, indicating high mathematical complexity. While it presents empirical tests on S&P 500 options and discusses computational constraints, the validation is limited by hardware limitations, leading to moderate empirical rigor.

Research Flowchart

  flowchart TD
  A["<b>Research Goal</b><br>American Option Pricing under<br>Time-Varying Rough Volatility"] --> B["<b>Key Methodology</b><br>Signature-Based Hybrid Framework"]
  
  B --> C["<b>Data Input</b><br>S&P 500 Equity-Index Options<br>Rolling Volatility Windows"]
  
  C --> D["<b>Computational Process</b><br>1. Estimate H(t) via<br>Gradient Boosted Ensemble<br>2. Regime Switch: Rough Bergomi vs. Calibrated Heston<br>3. RFF Acceleration for<br>Signature-Kernel Evaluation"]
  
  D --> E["<b>Key Findings/Outcomes</b><br>1. H(t) ≠ Constant: Persistent roughness violated<br>2. Adaptive Framework outperforms fixed-roughness baselines<br>3. Reduced duality gaps &<br>tractable real-time pricing"]

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 proposes a modular hybrid framework that extends signature methods with dynamic Hurst estimation, regime‑switching between rough Bergomi and Heston simulators, and accelerated kernel evaluation via Random Fourier Features for American option pricing.

Source sentence
We introduce a modular framework that extends the signature method to handle American option pricing under evolving volatility roughness.

Assumes 6

  • Constant risk‑free interest rate (approximately 4.5%). model
    Source sentence
    r: Constant risk-free interest rate (currently approximately 4.5%)
  • Pricing is performed under the risk‑neutral probability measure Q. model
    Source sentence
    Q: Risk-neutral probability measure under which expectations are taken
  • When the average forecasted Hurst parameter is below 0.5 the contract is priced with the rough Bergomi engine; otherwise the Heston engine is used. model
    Source sentence
    when the average H(t) is below 0.5, the contract is priced with the rough Bergomi engine; if it is at or above 0.5, we switch to a calibrated Heston engine
  • The gradient‑boosted tree ensemble can forecast H(t) from rolling windows of volatility signatures. data
    Source sentence
    A gradient-boosted tree ensemble is trained on sequential rolling windows of volatility signatures
Show 2 more
  • Signature‑kernel evaluations can be approximated with Random Fourier Features without significant loss of accuracy. model
    Source sentence
    this approximates the kernel in linear time and enables our model to handle longer volatility paths (N = 252) without significant loss in accuracy
  • Option data are daily at‑the‑money (ATM) contracts used for calibration. data
    Source sentence
    key parameters are computed from daily at-the-money (ATM) options data instead of relying on fixed calibration

Data

  • OptionMetrics options contract data · OptionMetrics via WRDS options commercial
    S&P 500 constituents (e.g., AAPL, META)
    Source sentence
    Data was sourced from OptionMetrics via WRDS (OptionMetrics 2025a). Two representative examples selected were two 10-day American put options (AAPL and META) from August 31, 2023.
  • Historical equity price data · OptionMetrics via WRDS market-prices commercial
    S&P 500 constituents
    Source sentence
    a second dataset containing historical equity prices in the format (date, ticker, close, return) (OptionMetrics 2025b)

Recognised sources: OptionMetrics IvyDB WRDS

Costs

Not applicable The paper reports no trading or execution result.

Finds 4

  • The hybrid framework improves pricing performance over fixed‑roughness baselines and reduces duality gaps in regimes where roughness changes. performance
    Source sentence
    The proposed hybrid framework provides a flexible structure that adapts to changing volatility roughness, improving performance over fixed-roughness baselines and reducing duality gaps in some regimes.
  • Pricing with the Deep Kernel model yields the narrowest primal‑dual gap (17.16%) for the AAPL put. pricing-calibration
    Source sentence
    Deep regressors improve alignment: both contain the premium, with the Deep Kernel model achieving the narrowest gap (17.16%).
  • When Ĥ > 0.5 the smooth Heston engine is selected; when Ĥ < 0.5 the rough Bergomi engine is selected. theoretical
    Source sentence
    For AAPL, the average Hurst forecast ¯Ht exceeded 0.5, prompting the use of the smooth Heston engine, whereas for META, the average Hurst forecast ¯Ht was less than 0.5, prompting the use of the Rough Bergomi engine.
  • Random Fourier Feature approximation enables linear‑time kernel computation while preserving accuracy for paths up to length 252. performance
    Source sentence
    This approximates the kernel in linear time and enables our model to handle longer volatility paths (N = 252) without significant loss in accuracy.

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 25 related papers). Coverage is partial, so this is neither replication nor refutation.

Cited by 2 papers in the archive (2 by independent authors): Volatility Modeling via EWMA-Driven Time-Dependent Hurst … (uses its method or data)

Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.