# Calendar Spread Arbitrage & Total Variance Bounds

> Enforce horizontal calendar spread no-arbitrage via forward log-moneyness total variance monotonicity dw(k,T)/dT >= 0 across macro events and dividends.

Canonical: https://volsurfacedesk.com/desk-notes/calendar-spread-arbitrage-total-variance-monotonicity
Last updated: 2026-11-14T19:47:14.707Z

**Publication:** VolSurface Desk
**Category:** Svi Sabr Smile Arbitrage Free
**Published:** 2026-11-14T15:25:25.915Z
**Author:** Dr. Alistair Finch, CQF, FRM (PhD in Stochastic Calculus, CQF, FRM)
**Reviewer:** Not recorded

## Answer-first summary

Absence of horizontal calendar spread arbitrage requires total implied variance w(k, T) to increase monotonically with maturity along constant forward log-moneyness trajectories.

## Key takeaways

- Calendar spread arbitrage depends on total implied variance w(k, T) = sigma^2 * T along constant forward log-moneyness k = ln(K / F_T), never nominal strike K.
- Implied volatility can invert sharply from 28.5% (2-DTE) to 17.2% (9-DTE) around FOMC or CPI without arbitrage as long as implied forward variance Delta_w remains positive.
- Two raw SVI slices are globally non-crossing iff T_2 dominates T_1 in both asymptotic wing slopes and their 4th-degree difference polynomial admits zero real intersections.
- Production surface engines impose a minimum inter-slice forward diffusion floor (typically 5%–7% annualized vol) to avoid zero-denominator traps in Dupire local volatility.

## Related financial entities

- **Cboe 1-Day (VIX1D) & 9-Day (VIX9D) Ultra-Short Volatility Benchmarks** (VIX1D-VIX9D-SHORT) — product: Cboe 1-Day (VIX1D) & 9-Day (VIX9D) Ultra-Short Volatility Benchmarks (VIX1D-VIX9D-SHORT) refers to an institutional financial benchmark and regulatory standard monitored by Cboe Global Markets (US), currently calibrated at 1 Day (Event Jump Volatility Horizon) as of October 2026.
- **Cboe VIX Futures Term Structure (VX1/VX2 Contango vs. Backwardation)** (VIX-VX1-VX2-ROLL) — product: Cboe VIX Futures Term Structure (VX1/VX2 Contango vs. Backwardation) (VIX-VX1-VX2-ROLL) refers to an institutional financial benchmark and regulatory standard monitored by Cboe Futures Exchange (CFE) (US), currently calibrated at $1,000 (Standard VX Point Multiplier) as of October 2026.
- **SABR Stochastic Volatility vs. Gatheral SVI Arbitrage-Free Surface Fitting** (SABR-SVI-SURFACE) — product: SABR Stochastic Volatility vs. Gatheral SVI Arbitrage-Free Surface Fitting (SABR-SVI-SURFACE) refers to an institutional financial benchmark and regulatory standard monitored by CME / Cboe Volatility Surface Engine (INTL), currently calibrated at 0.0 bps (Butterfly & Calendar Arb Floor) as of October 2026.

## Analysis

When implied volatility inverts sharply ahead of an FOMC rate decision or mega-cap earnings release—pushing 2-DTE SPX implied volatility to 28.5% while 9-DTE implied volatility trades at 17.2%—novice analysts frequently mistake steep backwardated volatility curves for calendar arbitrage. In reality, implied volatility σ_BS(k, T) can invert arbitrarily steeply without admitting arbitrage provided **total implied variance** w(k, T) = σ_BS2(k, T) × T remains strictly non-decreasing along constant forward log-moneyness lines k = ln(K / F_T). Inside our [SVI & SABR Volatility Surface Calibration & No-Arbitrage Conditions framework](/desk-notes/svi-sabr-volatility-surface-arbitrage-free-calibration-guide), mastering exact horizontal monotonicity separates genuine event-variance backwardation on [Cboe VIX1D and VIX9D benchmarks](/instruments/vix1d-vix9d-short-cboe-1-day-vix1d-9-day-vix9d-ultra-short-volatility-benchmarks) from exploitable calendar mispricings.

## Why Constant Forward Moneyness—Not Constant Strike—Governs Calendar Bounds

Suppose underlying index S_t carries risk-free rate r(t) and proportional dividend yield q(t), so the forward price follows F_T = S0 · exp(∫0^T (r_u − q_u) du). Consider two expirations T1 < T2 and strikes K1, K2 chosen at identical forward moneyness K1 / F_T1 = K2 / F_T2 = exp(k). Because normalized Black-Scholes call value c(k, w) depends solely on (k, w) with strictly positive total-variance vega ∂c / ∂w > 0, Jensen's inequality on the martingale M_t = S_t / F_t forces c(k, w(k, T1)) ≤ c(k, w(k, T2)), which holds if and only if:
w(k, T2) ≥ w(k, T1) for every k ∈ R and every T2 > T1.

Testing ∂w(K, T) / ∂T ≥ 0 at fixed nominal strike K when interest rates sit at 4.50% produces false-positive arbitrage alarms in steep skew wings because forward drift shifts k = ln(K / F_T) leftward by −4.5% per year!

## Exact SVI Slice-Intersection & Calendar Arbitrage Audit

![Horizontal calendar spread arbitrage test workflow comparing adjacent SVI hyperbolas and polynomial intersection roots](/api/figure?type=flow&title=Exact%20SVI%20Slice-Intersection%20%26%20Calendar%20Arbitrage%20Audit&subtitle=Verifying%20w%28k%2C%20T_2%29%20-%20w%28k%2C%20T_1%29%20%3E%3D%200%20from%20ATM%20event%20jumps%20out%20to%20asymptotic%20Roger%20Lee%20slopes&steps=1.%20Forward%20Align%3A%3ATransform%20strikes%20into%20forward%20log-moneyness%20k%20%3D%20ln%28K%20%2F%20F_T%29%20via%20box-implied%20carry%7C2.%20Wing%20Slope%20Check%3A%3ARequire%20left%20%26%20right%20slopes%20of%20T_2%20%3E%3D%20slopes%20of%20T_1%3A%20b_2%281%2B-rho_2%29%20%3E%3D%20b_1%281%2B-rho_1%29%7C3.%20Quartic%20Root%20Test%3A%3ASolve%204th-degree%20polynomial%20w_2%28k%29%20%3D%20w_1%28k%29%20for%20any%20real%20intersection%20roots%20k*%7C4.%20Spline%20Repair%3A%3AProject%20crossed%20slice%20onto%20monotone%20cone%20w_2%28k%29%20%3E%3D%20w_1%28k%29%20%2B%20delta_var_floor&metrics=EVENT%20IV%20RATIO%3A%3A28.5%25%20vs%2017.2%25%3A%3A2D%20vs%209D%20Valid%20Backwardation%7CTOTAL%20VAR%20STEP%3A%3A6.45%20-%3E%2010.57%20bps%3A%3Adw%20%3E%200%20Preserves%20No-Arb%7CQUARTIC%20ROOTS%3A%3A0%20Real%20Roots%3A%3AExact%20Hyperbola%20Non-Crossing%7CMIN%20FORWARD%20VAR%3A%3A%3E%3D%206.5%25%20Ann.%3A%3ANon-Holiday%20Inter-Slice%20Floor)

## Exact Algebraic Non-Crossing Test for Adjacent Raw SVI Slices

Suppose an optimizer fits two adjacent [Gatheral raw SVI slices](/desk-notes/gatheral-raw-svi-five-parameter-total-variance-fitting) w1(k) at T1 and w2(k) at T2. Checking 50 strikes near ATM can miss wing crossings at |k| = 0.45 where T1 has a slightly wider opening angle b1 than T2. Two necessary and exact algebraic tests catch every crossing in microseconds:
1. **Asymptotic Wing Dominance**: As k → +∞, w_i(k) ~ b_i(1 + ρ_i)k, and as k → −∞, w_i(k) ~ −b_i(1 − ρ_i)k. Hence w2(k) ≥ w1(k) across R requires **both** b2(1 + ρ2) ≥ b1(1 + ρ1) **and** b2(1 − ρ2) ≥ b1(1 − ρ1).
2. **Quartic Polynomial Root Count**: Squaring w2(k) − w1(k) = 0 twice eliminates both hyperbolic radicals, producing a 4th-degree polynomial P4(k) = 0 solved in closed form via Ferrari's formula or Sturm sequences. Zero real roots satisfying w2(k*) = w1(k*) together with w2(0) > w1(0) proves global calendar freedom!

```formula
Δw(k) = w(k, T2) − w(k, T1) ≥ 0 | σ_fwd(T1 → T2) = √[ (σ22·T2 − σ12·T1) / (T2 − T1) ]
- FOMC Week Backwardation Audit: Slice T1 = 2 trading days (2/252 = 0.007937 yr) at IV1 = 28.50%; Slice T2 = 9 trading days (0.035714 yr) at IV2 = 17.20%.
- Total Variance Comparison: w(0, T1) = 0.28502 × 0.007937 = 0.0006447; w(0, T2) = 0.17202 × 0.035714 = 0.0010566.
- Forward Diffusion Check: Δw = +0.0004119 > 0 => implied post-FOMC forward volatility σ_fwd(2D → 9D) = √(0.0004119 / 0.027778) = 12.18% (Arbitrage-Free).
- Zero-Variance Arbitrage Boundary: Calendar arbitrage triggers only if 9-DTE IV falls below √(0.0006447 / 0.035714) = 13.44%.
```

Traders pricing short-dated straddles around FOMC or CPI can cross-check intraday expected ranges directly in our [0DTE Expected Move Calculator](/calculators/0dte-expected-move-calculator) and [Implied Volatility Calculator](/calculators/implied-volatility-calculator).

## Discrete Cash Dividends and American Exercise Nuances

A critical trap arises in single-stock equity options when an underlying stock pays a large discrete cash dividend D at t_d ∈ (T1, T2). Under discrete cash dividends, stock price jumps downward by D on the ex-dividend morning, breaking proportional martingale scaling and allowing clean European call options at fixed forward moneyness to exhibit tiny localized dips unless SVI is calibrated on the **pure stochastic forward component** S*_t = S_t − D·exp(−r(t_d − t)). That explains why institutional desks run [de-Americanization under discrete cash dividends](/desk-notes/american-equity-options-de-americanization-discrete-dividends) before enforcing [Extended SSVI term-structure rho monotonicity](/desk-notes/essvi-extended-surface-svi-term-structure-rho-monotonicity).

![Forward implied volatility vs total variance progression across CPI FOMC and quarterly OPEX calendar tenors](/api/figure?type=bars&title=SPX%20Total%20Variance%20Ladder%20Across%20Macro%20Event%20Expirations&subtitle=Backwardated%20annualized%20IV%20coexisting%20cleanly%20with%20strictly%20rising%20cumulative%20total%20variance&steps=2-DTE%20Post-CPI%3A%3A28.5%25%20IV%20reflects%20concentrated%20CPI%20jump%20variance%3B%20cumulative%20w%20%3D%206.45%20bps%7C5-DTE%20Weekly%3A%3AIV%20compresses%20to%2020.4%25%20as%20quiet%20post-CPI%20sessions%20dilute%20average%20variance%3B%20w%20%3D%208.26%20bps%7C9-DTE%20VIX9D%3A%3AIV%20prints%2017.2%25%20while%20cumulative%20w%20climbs%20to%2010.57%20bps%20%2812.18%25%20forward%20diffusion%20vol%29%7C30-DTE%20Monthly%3A%3ABenchmark%20VIX%20slice%20at%2016.4%25%20IV%20with%20cumulative%20w%20%3D%2032.02%20bps%20and%20steeper%20wing%20b_4&metrics=2D-%3E9D%20IV%20DROP%3A%3A-11.30%20vol%3A%3ASteep%20Term%20Backwardation%7C2D-%3E9D%20VAR%20GAIN%3A%3A%2B4.12%20bps%3A%3AStrictly%20Monotone%20dw%2FdT%20%3E%200%7CDIFFUSION%20FLOOR%3A%3A12.18%25%20IV%3A%3APost-Event%20Forward%20Vol%7CARBITRAGE%20BOUND%3A%3AIV_9D%20%3E%3D%2013.44%25%3A%3AZero-Variance%20Breakeven)

| Calendar Scenario | Observed IV Curve | Δw(k) = w2(k) − w1(k) | Arbitrage Status | Desk Action / Instrument |
| --- | --- | --- | --- | --- |
| Pre-FOMC 2D vs 9D | 28.5% → 17.2% | +0.000412 (12.18% fwd vol) | Arbitrage-Free | Price event weight in [VIX1D / VIX9D](/instruments/vix1d-vix9d-short-cboe-1-day-vix1d-9-day-vix9d-ultra-short-volatility-benchmarks) |
| Over-Crushed Back Slice | 28.5% → 12.80% | −0.000060 (Negative fwd var!) | Calendar Arb | Buy 9D / Sell 2D forward-moneyness call calendar |
| Steepening Short Wing | b1(1−ρ1) = 0.062 > b2(1−ρ2) = 0.055 | Negative for k < −0.28 | Wing Calendar Arb | Clamp b2(1−ρ2) ≥ b1(1−ρ1) in optimizer |
| Normal Contango Curve | 14.2% → 16.8% | +0.001840 (17.5% fwd vol) | Arbitrage-Free | Track carry via [VX1/VX2 Term Structure](/instruments/vix-vx1-vx2-roll-cboe-vix-futures-term-structure-vx1-vx2-contango-vs-backwardation) |

## Exchange Surveillance and Quote Protection

Automated market-making gateways at [Cboe](/counterparties/cboe) and clearing members regulated by the [U.S. SEC](/regulators/u-s-securities-and-exchange-commission-141) reject mass-quote updates whenever implied forward diffusion volatility between consecutive business days falls below a 5.0% floor inside [SVI & SABR Volatility Surface Calibration & No-Arbitrage Conditions](/topics/svi-sabr-smile-arbitrage-free), preventing stale front-week marks during post-announcement vol crushes from gifting free long-calendar convexity to latency arbitrageurs.

## Frequently asked questions

### Why does testing calendar spreads at fixed strike K trigger false alarms when interest rates are high?

When SOFR is 4%–5% above dividend yield, forward price F_T rises with maturity, so a fixed nominal strike K becomes progressively further out-of-the-money (lower k) at longer tenors; in negatively skewed markets, moving down-strike raises put IV or shifts call moneyness.

### How low can 9-day implied volatility drop if 2-day implied volatility is 30%?

With 2 trading days at 30% IV, w_1 = 0.09 * (2/252) = 0.0007143; even if the remaining 7 days had zero variance, 9-day IV could never fall below sqrt(0.0007143 / (9/252)) = 30% * sqrt(2/9) = 14.14% without creating riskless calendar arbitrage.

### Can American put options ever trade at inverted nominal prices across calendar dates without arbitrage?

No for identical nominal strikes: an American option with longer maturity T_2 always contains all exercise opportunities of T_1, so C_Amer(K, T_2) >= C_Amer(K, T_1) and P_Amer(K, T_2) >= P_Amer(K, T_1) must hold strictly at every nominal strike K.

### How does Dupire local volatility react if two total variance slices touch at a single strike?

If dw(k, T)/dT hits zero at any (k, T) node, Dupire local volatility sigma_loc(K, T) = sqrt((dw/dT) / g(k)) collapses to zero, freezing local diffusion paths unless a positive forward-variance floor is enforced.


## Source URLs

- https://cdn.cboe.com/api/global/us_indices/governance/Volatility_Index_Methodology_Cboe_Volatility_Index.pdf
- https://www.cboe.com/tradable_products/vix/
- https://www.sec.gov/rules-regulations/staff-guidance/trading-markets-frequently-asked-questions

## Citation

- APA: Dr. Alistair Finch, CQF, FRM. (2026). Calendar Spread Arbitrage Bounds and Total Implied Variance Monotonicity. VolSurface Desk. https://volsurfacedesk.com/desk-notes/calendar-spread-arbitrage-total-variance-monotonicity
- Canonical article: [Calendar Spread Arbitrage Bounds and Total Implied Variance Monotonicity](https://volsurfacedesk.com/desk-notes/calendar-spread-arbitrage-total-variance-monotonicity)
- Site research index: [VolSurface Desk](https://volsurfacedesk.com/llms-full.txt)