Paper: arXiv 2611.25047
Authors: Martino Grasselli, Andrea Stanghellini, Sara Svaluto-Ferro
Abstract
Options on VXX bring the dynamics of the VIX futures curve into the joint SPX-VIX calibration problem. We construct a common model by specifying the VIX and then recovering instantaneous SPX variance from the requirement that its risk-neutral conditional expected average over the next thirty days equals the squared VIX. With truncated signatures of time-extended polynomial diffusions, this recovery reduces to finite-dimensional linear algebra and gives explicit access to the futures underlying VXX. The same specification therefore governs equity variance and rolling volatility exposure. To accommodate flexible specifications while respecting variance nonnegativity, we develop a regularized model and quantify the pricing errors of its tractable VIX and VXX proxies. Applications to short-dated option smiles illustrate joint calibration across the three markets, with Monte Carlo estimates of the proxy error bounds well below quoted bid-ask spreads.
Complexity vs Empirical Score
- Math Complexity: 9.0/10
- Empirical Rigor: 7.0/10
- Quadrant: Holy Grail — high math complexity, high empirical rigor
Why this score: This paper presents a highly complex mathematical framework using signature methods and polynomial diffusions for joint calibration. It demonstrates strong empirical application through calibration to market data and quantification of pricing errors, showing a robust blend of theory and practice. The novel approach to jointly model SPX, VIX, and VXX with explicit links is a significant contribution.
Research Flowchart
flowchart TD
A["Research Goal: Joint SPX, VIX, VXX Framework"] --> B{"Key Methodology: Common Model for SPX, VIX, VXX"};
B --> C["VIX Specification & SPX Variance Recovery"];
C --> D{"Computational Process: Truncated Signatures & Linear Algebra"};
D --> E["Inputs: Option Smiles (SPX, VIX, VXX)"];
E --> F{"Key Findings: Joint Calibration & Proxy Error Quantification"};
F --> G["Outcome: Flexible Model, Pricing Accuracy within Bid-Ask"];
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 joint model for SPX, VIX and VXX by specifying the VIX and recovering instantaneous SPX variance so that its risk‑neutral conditional expected average equals the squared VIX, using truncated signatures of time‑extended polynomial diffusions to obtain tractable VIX and VXX proxie
Source sentence
We construct a common model by specifying the VIX and then recovering instantaneous SPX variance from the requirement that its risk-neutral conditional expected average over the next thirty days equals the squared VIX. With truncated signatures of time-extended polynomial diffusions, this recovery reduces to finite-dimensional linear algebra and gives explicit access to the futures underlying VXX.
Assumes 6
- The underlying primary process X follows a polynomial diffusion as defined in (2.3). model
Source sentence
Let (Xt)t≥0 be a polynomial diffusion with dynamics dXt = b(Xt) dt + q a(Xt) dWt, X0 = x0 as in (2.3) and denote by bXt its time-extended signature at time t.
- The stock price dynamics follow a volatility‑based signature model under the risk‑neutral measure Q. model
Source sentence
We consider the following model for a stock price (St)t∈[0,T] under the risk-neutral measure Q. Definition 3.6. We say that a stochastic process (St)t∈[0,T] follows a volatility-based signature model if dSt = StσS t (ℓ) dBt for σS t (ℓ) = ⟨ℓ, bX≤n t ⟩ (3.4) and thus ST = S0 exp Z T 0 ⟨ℓ, bX≤n t ⟩dBt −1 2 Z T 0 ⟨ℓ, bX≤n t ⟩2 dt , for some ℓ∈Tn(Rd+1).
- Quantities used in the error bounds are integrable. statistical
Source sentence
Assume the hypotheses of Lemma 3.27 and that the quantities below are integrable.
- The VXX fee c is a fixed positive annual fee (c = 0.89%). market
Source sentence
In the present setting, however, the management fee c > 0 acts as a continuous negative carry on the VXX position. Hence, the absence of an early-exercise premium is no longer guaranteed.
Show 2 more
- Raw variance ⟨h(ℓ), bX≤2n t⟩ must be at least 10⁻⁶ at every calibration grid point through day 84. model
Source sentence
We will require the raw variance ⟨h(ℓ), bX≤2n t ⟩to be at least 10−6 at every calibration grid point through day 84.
- Early‑exercise premium for VXX American options is neglected because the fee is small and maturities are short. execution
Source sentence
Since the VXX options considered in our calibration have short maturities and the annual fee is small (according to the VXX’s prospectus, c = 0.89%), we neglect the early-exercise premium and use the corresponding European option price as an approximation: πAm VXX(T, K) ≈πVXX(T, K) := EQ e−rT e(r−c)TVXXT −K + .
Data
- Cboe DataShop option quotes options commercial
Source sentence
Historical bid and ask quotes for SPX, VIX, and VXX options dated 24 February 2026 were obtained through Cboe DataShop under license. The implied volatilities were computed by the authors from these quotes.
- Monte Carlo simulated paths synthetic unknown
Source sentence
Figure 3 reports implied-volatility smiles computed with 22 Figure 3: SPX, VIX and VXX implied volatilities after adding the slow-OU word (1, 1, 1) and the time word (0, 0, 0) to the complete level-two VIX functional and jointly recalibrating, using 106 independent Monte Carlo paths. The specification has 23 VIX coefficients. the calibrated parameters held fixed on 106 independent Monte Carlo trajectories.
Costs
Not applicable
Finds 5
- Joint calibration across SPX, VIX and VXX markets yields proxy pricing errors well below quoted bid‑ask spreads. pricing-calibration
Source sentence
Applications to short-dated option smiles illustrate joint calibration across the three markets, with Monte Carlo estimates of the proxy error bounds well below quoted bid-ask spreads.
- The SPX downside skew and right‑wing upturn are well reproduced after adding the slow‑OU and time words to the VIX functional. pricing-calibration
Source sentence
It shows that the SPX downside skew and right-wing upturn remain well reproduced. The VIX curve is steeper than in the level-two benchmark and the underestimation of the upper wing is reduced.
- Monte Carlo estimates of VIX and VXX call‑price errors introduced by the tractable proxies are smaller than market bid‑ask spreads. pricing-calibration
Source sentence
Finally, Table 2 reports Monte Carlo estimates of the VIX and VXX call-price errors introduced by working with tractable proxies, together with the corresponding market bid–ask spreads.
- The VXX implied‑volatility curve shifts upwards in the OU–CIR specification, with all model points lying within or at the boundaries of quoted bid‑ask bands. pricing-calibration
Source sentence
For VXX, the implied-volatility curve shifts upwards, with all displayed model points lying within or at the boundaries of the quoted bid–ask bands.
- The regularized model maintains variance non‑negativity while providing flexible specifications, with error bounds giving a financial interpretation of the tractability‑accuracy trade‑off. theoretical
Source sentence
To accommodate flexible specifications while respecting variance nonnegativity, we develop a regularized model and quantify the pricing errors of its tractable VIX and VXX proxies. The accompanying error bounds give this compromise a financial interpretation: the gain in tractability can be assessed against the pricing discrepancy it introduces.
Checked by others
No other paper in the archive has been found to support or contradict these findings (compared with 21 related papers). Coverage is partial, so this is neither replication nor refutation.
Code
No repository by the authors found in the paper.
Extracted 2026-11-20 · same card as JSON · A language model chose these items: check the source sentence before relying on one.