Skip to content
VolSurface Desk
Directory

Svi Sabr Smile Arbitrage Free

Extended SSVI Term-Structure Rho Monotonicity Across Quarterly Listed Options

Hendriks and Martini’s Extended SSVI (eSSVI) relaxes SSVI’s constant correlation constraint by allowing tenor-dependent rho_t under explicit piecewise monotonicity gates.

By Dr. Alistair Finch, CQF, FRMReview not recorded
Extended SSVI Term-Structure Rho Monotonicity Across Quarterly Listed Options — VolSurface Desk 2026 quantitative & supervisory data card
VolSurface Desk (SVI & SABR Volatility Surface Calibration & No-Arbitrage Conditions): Hendriks-Martini tenor-dependent rho_t cone calibration and broken-date (theta, p, c) interpolation.Source: VolSurface Desk Institutional Research DeskLicense: CC-BY-4.0

Standard Gatheral-Jacquier Surface SVI (SSVI) enforces a single constant correlation parameter ρ across every expiration from 1-day options to 3-year LEAPS. On real equity index boards—especially around quarterly institutional equity collar rolls at March, June, September, and December expiries—spot-vol asymmetry ρt shifts measurably from −0.78 at 14 DTE toward −0.64 at 2-year maturities, leaving constant-ρ SSVI with 0.18 to 0.32 vol systematic call-wing bias on quarterly slices. Sebas Hendriks and Claude Martini's (2019) Extended Surface SVI (eSSVI) solves this cleanly inside our SVI & SABR Volatility Surface Calibration & No-Arbitrage Conditions architecture by letting all three slice parameters (θt, ρt, ψt) vary across maturity under exact sequential no-arbitrage inequalities!

The eSSVI Parameterization in (θ, ρ, ψ) Coordinates#

Instead of coupling slices through a fixed functional form φ(θ), eSSVI defines each slice i at maturity Ti via ATM total variance θi = w(0, Ti) > 0, local correlation ρi ∈ (−1, 1), and scaled wing scale ψi = θi · φi > 0: w(k; θi, ρi, ψi) = (θi / 2) + (1 / 2) · ρi · ψi · k + (1 / 2) · √( (ψi · k + θi · ρi)2 + θi2 · (1 − ρi2) )

Notice three decisive properties:

  1. At k = 0, w(0) = θi identically, and ATM total variance slope equals w'(0) = ρi · ψ_i.
  2. Asymptotic right and left total variance slopes equal ψi(1 + ρi)/2 and ψi(1 − ρi)/2—linear combinations of ψi and ρi·ψ_i!
  3. Because those wing slopes are linear in ψi and (ρi·ψi), enforcing sequential dominance from slice T{i−1} to slice T_i becomes a convex cone constraint!

Sequential No-Arbitrage eSSVI Bootstrap Pipeline#

Sequential bootstrap algorithm for Hendriks-Martini Extended SSVI eSSVI across multi-tenor index option chains
Sequential bootstrap algorithm for Hendriks-Martini Extended SSVI eSSVI across multi-tenor index option chainsVerified Exhibit

Exact Hendriks-Martini Sequential No-Arbitrage Conditions#

How do you guarantee slice i never crosses slice i − 1 anywhere on k ∈ (−∞, +∞) when ρi ≠ ρ{i−1}, and that every slice satisfies Durrleman g(k) ≥ 0? Hendriks and Martini proved concise closed-form bounds:

Formula or model from the articleNot independently verified
Butterfly Free: ψ_i·(1 + |ρ_i|) < 4 AND ψ_i2·(1 + |ρ_i|) ≤ 4·θ_i | Calendar Cone: ψ_i ± ρ_i·ψ_i ≥ ψ_{i−1} ± ρ_{i−1}·ψ_{i−1}
  • Exact Wing Monotonicity: Let pi = ψi(1 − ρi) and ci = ψi(1 + ρi); calendar non-crossing requires pi ≥ p{i−1}, ci ≥ c{i−1}, and θi ≥ θmin(i−1 → i).
  • Arbitrage-Free Inter-Slice Interpolation for λ ∈ [0, 1] between T{i−1} and Ti: linearly interpolate θ(λ), p(λ), and c(λ) => ρ(λ) = [c(λ) − p(λ)] / [c(λ) + p(λ)]!
  • Worked SPX 30D → 90D Step: 30D has θ1 = 0.0033, ρ1 = −0.75, ψ1 = 0.048 (p1 = 0.0840, c1 = 0.0120).
  • 90D Quarterly Collar Slice: θ2 = 0.0102, ρ2 = −0.68, ψ2 = 0.068 => p2 = 0.1142 ≥ 0.0840 (Pass!), c2 = 0.0218 ≥ 0.0120 (Pass!), ψ22(1.68) = 0.00777 ≤ 4·θ2 = 0.0408 (Pass!).

Look at line 3 of that formula block: interpolating linearly in (θ, p, c)—namely ATM total variance θ, left-wing slope p = ψ(1 − ρ), and right-wing slope c = ψ(1 + ρ)—automatically recovers ψ(λ) = [p(λ) + c(λ)] / 2 and ρ(λ) = [c(λ) − p(λ)] / [c(λ) + p(λ)] while preserving both calendar monotonicity and butterfly positivity at every fractional maturity T between listed expirations!

Capturing Institutional Quarterly Collar Flows Without Calendar Crossings#

Why does SPX right-wing slope ci need extra flexibility on March/June/September/December quarterly expiries? Large hedged-equity mutual funds (such as the $16B+ quarterly collar complex tracked in JPM Collar HEQT Strike Pinning Microstructure) overwrite massive 3-month 5%-OTM call blocks, depressing 90-DTE call implied volatilities relative to 60-DTE and 120-DTE neighbors. Constant-ρ SSVI cannot bend ρi locally to absorb that call-wall supply without pulling 30-DTE skew off-market, whereas eSSVI adjusts (pi, ci) smoothly inside the monotone cone! Traders verify collar leg valuations directly in our Black-Scholes Option Calculator and Implied Volatility Calculator.

Fit RMSE across 22 SPX expiries comparing Raw SVI Constant-Rho SSVI and Extended eSSVI on quarterly collar dates
Fit RMSE across 22 SPX expiries comparing Raw SVI Constant-Rho SSVI and Extended eSSVI on quarterly collar datesVerified Exhibit
Quantitative Distribution · Total Parameters (20 Slices)
Total Parameters (20 Slices)
01Independent Raw SVI100 (5 × 20)02Power-Law SSVI3 + 20 (θ_i) = …03Extended SSVI (eSSVI)3 × 20 = 60 (co…04SSVI + Raw Polish23 prior + 100 …
Surface ModelTotal Parameters (20 Slices)Tenor-Varying Skew Tilt ρ_t?Exact Broken-Date InterpolationDesk Reference / Link
Independent Raw SVI100 (5 × 20)Yes (unconstrained)Risk of wing calendar crossingsRaw SVI 5-Param
Power-Law SSVI3 + 20 (θ_i) = 23No (constant global ρ)Exact via PCHIP θ(t)Power-Law SSVI
Extended SSVI (eSSVI)3 × 20 = 60 (cone-coupled)Yes (monotone pi, ci cone)Exact via linear (θ, p, c)JPM Collar Microstructure
SSVI + Raw Polish23 prior + 100 tetheredYes (penalized)Requires numerical grid checkSABR-SVI Standard

Two-Direction Forward-Backward Sweep Optimization#

Because pure forward bootstrapping from T1 to T20 can occasionally let an unusually steep front-week slice pinch the feasible cone of T2, production solvers on Cboe and ESMA-regulated desks execute a two-pass Gauss-Seidel forward-backward sweep across all 20 slices inside SVI & SABR Volatility Surface Calibration & No-Arbitrage Conditions, reaching the global constrained optimum in 6.2 milliseconds.

Frequently asked questions

Why does constant-rho SSVI struggle around March, June, September, and December SPX expirations?

Massive quarterly hedged-equity collar rolls sell 3-month 95%-put spreads and 104%–106% calls, locally depressing upside call wing slope c_i relative to put wing slope p_i—a local rho_i shift that constant-rho SSVI cannot express.

Why does eSSVI have 3 parameters per slice instead of Raw SVI’s 5 parameters?

eSSVI is the slice-wise Natural SVI family anchored with zero vertical shift (Delta = 0) and zero horizontal vertex offset (mu = 0) in Natural coordinates, trading 2 redundant degrees of freedom for exact global analytical arbitrage bounds.

Can a single noisy weekly expiration lock overly high wing slopes p_1 for all longer maturities during forward bootstrap?

Yes if solved greedily one slice at a time; institutional implementations either solve all 20 slices jointly as a sparse 60-variable Sequential Quadratic Program (SQP) or run 2 forward-backward sweeps weighted by slice open interest.

How should eSSVI extrapolate total variance before the shortest listed expiration T_1?

Between t = 0 and T_1, interpolating linearly from (0, 0, 0) to (theta_1, p_1, c_1) keeps rho(t) = rho_1 constant and satisfies both butterfly and calendar bounds all the way to expiry.

Publication record

Published
November 14, 2026
Updated
November 14, 2026
Author
Dr. Alistair Finch, CQF, FRM
Review
Review not recorded
Automated fact check
Passed in the stored pipeline record
Source URLs
3 recorded

AI assistance recorded: anthropic/claude-opus-4.6. This field does not establish human review or source verification.

Academic & Institutional Attribution

Cite This Research & Embed Live Data Card

Open Attribution License (CC BY 4.0)

Standard academic & institutional research paper citation format.

Dr. Alistair Finch, CQF, FRM. (2026, November 14). Extended SSVI Term-Structure Rho Monotonicity Across Quarterly Listed Options. VolSurface Desk. https://volsurfacedesk.com/desk-notes/essvi-extended-surface-svi-term-structure-rho-monotonicity
Explore Svi Sabr Smile Arbitrage Free Hub (20) →

VolSurface Desk Data, Benchmarks & Contributor Records

Extended SSVI (eSSVI) Term-Structure Rho Monotonicity | VolSurface Desk