Paper: arXiv 2406.16131

Authors: Peter K. Friz, Jim Gatheral

Abstract

The skew-stickiness-ratio (SSR), examined in detail by Bergomi in his book, is critically important to options traders, especially market makers. We present a model-free expression for the SSR in terms of the characteristic function. In the diffusion setting, it is well-known that the short-term limit of the SSR is 2; a corollary of our results is that this limit is \(H+3/2\) where \(H\) is the Hurst exponent of the volatility process. The general formula for the SSR simplifies and becomes particularly tractable in the affine forward variance case. We explain the qualitative behavior of the SSR with respect to the shape of the forward variance curve, and thus also path-dependence of the SSR.

Complexity vs Empirical Score

  • Math Complexity: 8.5/10
  • Empirical Rigor: 3.0/10
  • Quadrant: Lab Rats — high math complexity, limited empirical testing

Why this score: The paper introduces highly advanced stochastic calculus and characteristic function techniques to derive a model-free expression for the SSR, with heavy reliance on derivatives and integrals. However, it lacks any backtesting, code, or implementation details, focusing purely on theoretical derivation without empirical data.

Research Flowchart

  flowchart TD
  A["Research Goal:<br>Derive Model-Free Expression<br>for Skew-Stickiness-Ratio (SSR)"] --> B["Methodology:<br>Analysis via Characteristic Function"]
  B --> C["Key Input:<br>Affine Forward Variance Model"]
  C --> D{"Computational Process"}
  D --> E["General Formula Derivation"]
  D --> F["Short-Term Limit Analysis"]
  E --> G["Key Findings/Outcomes"]
  F --> G
  G --> H["Limit: H + 3/2<br>(H = Hurst Exponent)"]
  G --> I["SSR Behavior<br>(Path-dependent & Shaped by<br>Forward Variance Curve)"]
  style A fill:#f9f,stroke:#333,stroke-width:2px
  style G fill:#ccf,stroke:#333,stroke-width:2px

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 provides a generic, essentially model‑free expression for the skew‑stickiness‑ratio (SSR) in terms of the characteristic function, and an explicit formula for the SSR in affine forward variance (AFV) models via a convolution Riccati integral equation.

Source sentence
Firstly, Proposition 3.1gives a generic and essentially model-free expression for the SSR in terms of the characteristic function. Secondly, Proposition 4.1 gives an explicit formula for the SSR in affine forward variance (AFV) models, in terms of the solution of a convolution Riccati integral equation.

Assumes 6

  • The forward variance process ξ_t(u) is adapted to the filtration generated by the Brownian motion W. model
    Source sentence
    If, as is usual in stochastic volatility modeling, we assume that ξ is adapted to the filtration generated by W, the ATM volatility σt(T) can only depend on the volatility state variables, namely the forward variances {ξt(u) : u > t}.
  • The model (2.1) is scale‑invariant and option prices can be computed under this forward variance specification. model
    Source sentence
    In particular, such a model is scale-invariant, with ξ adapted to the filtration generated by W. Assuming that we can compute option prices in model (2.1), computing St(T) is in principle straightforward.
  • For short‑time analysis the forward variance curve is assumed flat (ξ_t(u)=̅V). model
    Source sentence
    If the forward variance curve is flat with ξt(u) = ¯V, the SSR does not depend on the level V at all!
  • The volatility process follows a stochastic volatility dynamics with Hurst exponent H (rough volatility). model
    Source sentence
    In the diffusion setting, it is well-known that the short-term limit of the SSR is 2; a corollary of our results is that this limit is $H+3/2$ where $H$ is the Hurst exponent of the volatility process. 14 In Figure 5.1, with ξ = 0.025, ν = 0.4, and ρ = −.8, we plot the SSR for various values of H, numerically using (4.2) and using the forest formula (5.6) with the above substitu-tions.
Show 2 more
  • The ATM volatility σ_t(T) depends only on the forward variances {ξ_t(u): u>t}. model
    Source sentence
    If, as is usual in stochastic volatility modeling, we assume that ξ is adapted to the filtration generated by W, the ATM volatility σt(T) can only depend on the volatility state variables, namely the forward variances {ξt(u) : u > t}.
  • In AFV models the diffusion coefficient satisfies ft(ξ)=√V_t. model
    Source sentence
    With φ = eψ, the result follows. □ 4 The SSR in affine forward variance (AFV) models Recall from [GKR19] that in AFV models, dξt(u) = κ(u −t) √ Vt dWt, (4.1) so ft(ξ) = √Vt.

Data

No dataset named in the text.

Costs

Not applicable The paper reports no trading or execution result.

Finds 5

  • Proposition 3.1 gives a generic, model‑free expression for the SSR in terms of the characteristic function. theoretical
    Source sentence
    Firstly, Proposition 3.1gives a generic and essentially model-free expression for the SSR in terms of the characteristic function.
  • Proposition 4.1 provides an explicit SSR formula for affine forward variance models via a convolution Riccati integral equation. theoretical
    Source sentence
    Secondly, Proposition 4.1 gives an explicit formula for the SSR in affine forward variance (AFV) models, in terms of the solution of a convolution Riccati integral equation.
  • Lemma 5.4, using the forest expansion, characterizes the short‑time behavior of the SSR. theoretical
    Source sentence
    By using the forest expansion of the characteristic function, we were able to derive a model-free characterization of the short-time behavior of the SSR in Lemma 5.4.
  • Corollary 5.6 confirms prior formal claims on the limiting behavior of the SSR, including the short‑time limit H+3/2. theoretical
    Source sentence
    Corollary 5.6 then confirms various prior formal claims on the limiting behavior of the SSR.
  • Observed SSR values in markets are typically between 1.0 and 1.3, which appear inconsistent with the affine model computations shown in Figure 7.2. negative
    Source sentence
    It is important to highlight that observed skew stickiness ratios are typically between 1.0 and 1.3, seemingly inconsistent with our computations in affine models as illustrated in Figure (7.2).

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 17 related papers). Coverage is partial, so this is neither replication nor refutation.

Cited by 2 papers in the archive (2 by independent authors)

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.