How CME Futures-and-Options Margin Algorithms Turn Portfolios into Performance Bonds: SPAN Risk Arrays, Spread Credits, Short-Option Minimums and SPAN 2

Reader question: A futures portfolio can contain outright positions, calendar spreads, options, correlated commodities and deep out-of-the-money short options. Why is margin not simply “10% of notional,” and how can a clearing system recognize genuine offsets without giving away too much credit?

The answer is portfolio risk simulation. CME’s legacy SPAN methodology groups related positions into combined commodities, reprices them across standardized price-and-volatility scenarios, finds the worst scanning loss, adds charges for risks that the simple scenarios do not fully capture, recognizes specified spread offsets, and applies safeguards such as the short-option minimum. CME’s newer SPAN 2 framework extends the architecture with a more modern risk model and additional components. Current CME futures-and-options portfolios can therefore be margined under SPAN and/or SPAN 2 depending on product and implementation.

The computational idea is more important than the brand name: margin is a conservative portfolio loss estimate under defined stress and offset rules, not a percentage copied from gross position size.

What this page owns — and what it does not

This page owns:

cleared futures/options positions + margin parameters → risk scenarios + spread treatment + safeguards → portfolio performance-bond requirement.

It does not replace ISDA SIMM, which is a sensitivity-based initial-margin framework for uncleared derivatives; CCP default waterfalls; or variation-margin calls. This page owns the futures/options portfolio margin calculation itself.

This is public clearing-risk mathematics, not advice on how much leverage to use or which derivatives to trade.

Why gross notional is a weak margin measure

Suppose a trader is long one December futures contract and short one March contract on the same underlying.

Gross notional can be very large:

|Long Notional| + |Short Notional|.

But much of the broad market-direction risk offsets. The real residual risk is the calendar spread: December and March prices can move differently.

A flat percentage of gross notional ignores the offset. A flat percentage of net notional can ignore the basis risk. Portfolio margin tries to measure both.

Performance bond versus option premium

CME commonly calls futures margin a performance bond. It is collateral supporting the participant’s ability to meet obligations, not a down payment that buys part of the futures contract.

Option premium is different. A long option buyer pays premium and cannot ordinarily lose more than that paid premium on the option itself. A short option writer can face much larger losses and therefore requires margin beyond simply retaining the premium.

A margin engine must distinguish:

premium value state ≠ portfolio risk requirement.

Legacy SPAN: group positions by ultimate underlying

CME’s SPAN methodology groups related instruments into combined commodities.

A combined commodity can include:

• futures across expiries;
• calls and puts;
• different strikes;
• other eligible instruments;

that share the same ultimate underlying.

This grouping lets the engine recognize that a call, put and futures position on the same underlying are economically related even though they are separate contracts.

Step 1: build a risk array for each contract

SPAN represents each instrument’s profit/loss response across a standardized set of risk scenarios.

For options, CME’s public material describes 16 scenarios combining changes in underlying price and implied volatility, including extreme scan scenarios.

Conceptually, for instrument i and scenario s:

P/L_i,s = RevaluedPrice_i,s − CurrentPrice_i.

For a portfolio:

Portfolio P/L_s = Σ Position_i × P/L_i,s.

The risk array stores these scenario responses so firms can reproduce the clearing calculation from published risk-parameter files.

Step 2: scan for the worst portfolio loss

For one combined commodity:

Scan Risk = max_s(−Portfolio P/L_s).

The worst loss across the specified scenarios becomes the core directional/volatility risk charge.

This is a scenario-based Value-at-Risk-like architecture: it asks how badly the portfolio behaves under a defined grid of market shocks.

A simple futures example

Suppose one futures contract gains or loses $50 for each index point and the scan range considers a 20-point adverse move.

One long contract could produce a $1,000 loss in that scenario:

20 × $50 = $1,000.

If the portfolio also holds an offsetting option that gains $300 under the same scenario, combined scan loss becomes $700 rather than $1,000.

Portfolio offsets are recognized scenario by scenario rather than through one static hedge ratio.

Why options need both price and volatility shocks

An option’s value is nonlinear.

A short call can lose because:

• the underlying price rises;
• implied volatility rises;
• both occur together.

A margin grid that shocks only underlying price can miss volatility risk.

The 16-scenario SPAN architecture therefore includes volatility changes as well as price changes.

Step 3: intra-commodity spread charge

Calendar spreads reduce directional risk but create basis risk.

December and March futures on the same commodity are highly related, not identical.

SPAN can therefore add an intra-commodity spread charge after scanning to reflect residual risk between contract months.

A perfect one-for-one long/short calendar spread should not automatically receive zero margin.

Step 4: delivery risk

Some physically delivered futures become riskier near delivery because positions can create concentration, timing or settlement exposures not captured well by ordinary price scans.

SPAN can apply delivery-related charges.

This is another reason a portfolio with the same delta can have different margin depending on contract month.

Step 5: inter-commodity spread credits

Different but correlated products can offset risk.

For example, certain interest-rate contracts or agricultural products can have economically related price moves.

SPAN can grant an inter-commodity spread credit for approved relationships.

The important word is approved. A trader cannot invent a correlation from one month of historical data and expect the clearing system to accept it.

Combined-commodity requirement

CME’s public SPAN overview describes the broad structure as:

Scan Risk + Intra-Commodity Spread Charge + Delivery Risk − Inter-Commodity Spread Credit.

The result is then compared with another safeguard: the short-option minimum.

Short-option minimum: protect against scenario-grid blind spots

A deep out-of-the-money short option can show little or no loss under an ordinary scenario grid because the option remains out of the money in most shocks.

But if the underlying makes a larger move, the option can become valuable to its holder and dangerous to the short seller.

CME therefore applies a Short Option Minimum (SOM).

For the relevant combined commodity:

SPAN Risk Requirement = max(Scenario/Spread Requirement, Short Option Minimum).

The SOM is a floor against false comfort from a finite scenario set.

Counterexample: zero scan risk does not mean zero short-option risk

Suppose a short call is so far out of the money that all 16 standard scenarios leave it nearly worthless.

Scan risk can be tiny.

The short-option minimum can still create a positive margin requirement because losses beyond the scenario grid remain possible.

This is a direct example of a model-residual add-on.

Option value affects the account calculation too

CME’s public options-margin explanation notes that long option value can create a credit and short option value a debit in PC-SPAN-style calculations.

This reflects the economic fact that:

• a long option holder has already paid value for an asset;
• a short option writer has received premium but still owes the contingent payoff.

The exact account requirement can therefore include both risk requirement and option-value treatment rather than simply the worst scenario loss.

Portfolio aggregation across currencies

CME’s SPAN overview describes calculating requirements across combined commodities, converting them to a common currency and summing them into a portfolio risk figure.

This creates another data dependency:

risk currency → conversion rate → base margin currency.

Stale FX conversion can distort total margin even when every product-level calculation is correct.

Why spread credits cannot be unlimited

Correlations break during stress.

Two contracts that usually move together can temporarily diverge because of:

• delivery conditions;
• location differences;
• quality differences;
• policy events;
• liquidity shocks.

Clearing spread credits are therefore calibrated conservatively and can be changed when relationships weaken.

A stylised combined-commodity example

Suppose a portfolio has:

• scan risk = $8,000;
• intra-commodity spread charge = $1,500;
• delivery risk = $500;
• inter-commodity spread credit = $2,000.

Before the short-option floor:

8,000 + 1,500 + 500 − 2,000 = $8,000.

If SOM = $6,000, the scenario/spread requirement of $8,000 wins.

If SOM = $10,000, the margin floor becomes $10,000.

The numbers are illustrative; current product parameters come from CME risk files and clearing notices.

SPAN is parameter-driven

The methodology is only half the algorithm. The current risk parameter file supplies product-specific settings such as scan ranges, volatility ranges, spread parameters and short-option minimums.

Therefore:

same positions + different effective-date parameters → different margin.

Reproducibility requires storing the exact parameter-file version.

Risk arrays make independent replication possible

CME publishes SPAN reference documents and parameter files to clearing firms and market participants.

A firm can load positions and the same risk arrays into approved tools such as CME CORE or SPAN software and compare the result with clearing reports.

This is a powerful control because margin is not a black-box number with no audit trail.

SPAN 2: do not assume every current product still uses the legacy 16-scenario engine

CME has been rolling out SPAN 2, a newer portfolio-margin framework. CME’s current rollout material states that production tools can support both SPAN and SPAN 2 methodologies.

The architectural lesson is:

margin_methodology = versioned product attribute.

A production engine must know whether a portfolio/product is under legacy SPAN, SPAN 2 or another approved treatment rather than blindly applying the 1988-era formula to every CME position.

SPAN 2 uses a richer risk architecture

CME describes SPAN 2 as a modernized framework with risk components and parameterization designed to reflect current portfolio risks more directly.

Public CME notices refer to subcomponents including:

• historical value-at-risk;
• stress value-at-risk;
• volatility floors;
• product-group parameterization;
• additional liquidity/concentration-style risk treatment where applicable.

The exact component set can depend on the relevant product pod and current methodology.

Current 2026 parameter changes prove the model is live, not static

CME published an August 2026 advisory on SPAN 2 equity-model parameter changes. The notice described changing a volatility floor in the historical-VaR subcomponent and modifying stress-scenario parameters, with portfolio margin impacts depending on portfolio risk profile.

This is why a current margin article should not publish one timeless parameter table and call it “the SPAN 2 formula.”

The framework is governed and recalibrated.

Coverage targets and anti-procyclicality

CME’s current futures-and-options margin-model page lists minimum coverage levels of 99% for major product groups and describes measures such as volatility floors and a benchmark margin buffer.

The buffer is intended to reduce frequent procyclical margin changes during periods of rising volatility.

This exposes a fundamental tension:

• margin must rise when risk rises;
• margin should not amplify stress unnecessarily through abrupt mechanical jumps.

Margin is not a prediction of the maximum possible loss

A 99% coverage target does not mean losses cannot exceed margin.

Extreme market gaps, liquidity shocks and model misspecification can produce losses beyond the initial margin amount.

That residual risk is why CCPs also maintain default funds, capital and recovery resources, as explained in the CCP default-waterfall article.

Counterexample: a hedged portfolio can require more margin after volatility rises

Suppose a delta-neutral options portfolio is well hedged for small price changes.

If implied volatility jumps, option values can move substantially and the clearing model can widen scenario losses or recalibrate parameters.

“Delta neutral” does not mean “margin stable.”

Counterexample: two opposite futures are not riskless

Long December and short March futures net to zero gross direction but retain calendar-spread risk.

An intra-commodity spread charge captures that residual.

Counterexample: correlation credit can disappear when stress changes the relationship

A portfolio may receive an intercommodity offset during normal conditions.

If clearing parameters are recalibrated after correlations weaken, margin can rise even though positions are unchanged.

The parameter set is part of the risk state.

Counterexample: long and short options with the same notional are asymmetric

A long option’s loss is generally limited to premium paid.

A short option’s loss can be far larger.

Margin must therefore depend on payoff shape, not only notional.

Inputs and outputs

A futures/options margin engine can require:

• clearing account and positions;
• contract identifiers, expiries and strikes;
• current settlement prices;
• option volatility inputs;
• SPAN or SPAN 2 methodology identifier;
• current risk parameter file;
• scan ranges/scenario parameters;
• spread relationships and credits;
• short-option-minimum parameters;
• delivery/concentration parameters;
• currency conversion rates;
• effective date.

Outputs can include:

• scenario losses;
• scan risk;
• spread charges and credits;
• short-option minimum;
• option-value adjustment;
• combined-commodity requirement;
• portfolio performance bond;
• margin attribution by product/risk component.

Evidence polarity: what supports confidence?

Evidence for a correct margin calculation includes positions reconciling to clearing records, risk arrays loaded from the correct effective-date file, the worst scenario independently identified, spread credits matching approved relationships, SOM applied to deep short options, and final requirements reconciling to CME CORE or clearing reports.

Evidence against confidence includes a hedged calendar spread receiving zero margin, deep out-of-the-money short options receiving no floor, an obsolete SPAN file remaining in use after parameter changes, SPAN 2 products forced through legacy 16-scenario logic, or margin results that cannot be attributed to scenario/spread/add-on components.

Weak links in implementation

position-sign error. Long and short positions are reversed.

combined-commodity mapping error. Related contracts fail to offset or unrelated contracts receive false offsets.

scenario staleness. Old scan parameters survive a clearing update.

spread-credit overreach. Correlation credits are applied outside the approved relationship.

SOM omission. Deep short options are treated as zero risk.

delivery-risk omission. Near-delivery concentration is ignored.

methodology-version error. SPAN and SPAN 2 are confused.

currency-conversion error. Multi-currency requirements use stale FX rates.

Diagnostics: how to test the engine

• 16-scenario replay: for a legacy-SPAN option portfolio, reproduce all scenario P/L values and identify the same worst scan loss.
• calendar-spread test: long/short adjacent futures should retain non-zero spread risk.
• intercommodity-credit test: apply an approved offset, then remove it and verify margin rises.
• SOM test: deep OTM short options must retain positive minimum risk where applicable.
• long-option test: ensure long option premium/value treatment does not create unlimited loss.
• parameter-version test: switch effective-date risk files and verify margin changes where CME parameters changed.
• SPAN/SPAN2 routing test: each product maps to the correct current methodology.
• currency test: convert product-level requirements to a common currency consistently.
• CME CORE reconciliation: independently compare a test portfolio with CME’s supported margin output.
• stress exceedance test: show that losses can exceed initial margin under shocks beyond the model envelope and route residual risk to CCP stress/default-fund analysis.

What would falsify confidence?

Confidence should be withdrawn if current clearing reports cannot be reproduced; if parameter provenance is missing; if the methodology identifier is unknown; if offsets exceed approved relationships; if short-option safeguards disappear; or if a CME model advisory changes parameters without affecting portfolios that should be in scope.

Alternatives and limits

ISDA SIMM uses sensitivities and prescribed correlations for uncleared derivatives. Historical-simulation VaR and expected shortfall are used in other risk contexts. Brokerage firms can also impose house margin above clearing minimums.

CME clearing margin is therefore not necessarily the final margin a customer sees. Customer-level requirements can include broker add-ons, concentration charges, regulatory overlays or account-specific policies.

How this connects to the surrounding knowledge estate

ISDA SIMM provides a contrasting sensitivity-based margin framework. Variation margin handles current mark-to-market transfer. CCP waterfalls absorb losses beyond defaulter margin. SPAN/SPAN 2 own the portfolio-loss-to-performance-bond transformation for cleared futures/options.

Verification and update triggers

Preserve methodology version, product scope, risk parameter file, position snapshot, scenario set, spread mapping, SOM parameters, delivery charges, currency rates and clearing date. Revalidate after CME clearing advisories, SPAN 2 migrations, volatility regime shifts, contract launches, spread-parameter changes or any margin reconciliation break.

Primary and high-quality references

• CME Group, CME SPAN: Standard Portfolio Analysis of Risk.
• CME Group, SPAN Methodology Overview, including combined commodities and spread treatment.
• CME Group, Margins on Options, including the 16 risk scenarios and short-option minimum.
• CME Group, Futures and Options Margin Model, including coverage targets, volatility floors and benchmark margin buffers.
• CME Group, SPAN 2 Framework Rollout.
• CME Group, SPAN 2 Framework Equity Model Parameter Changes — effective August 18, 2026.

Educational boundary: This article explains public futures/options margin methodology. It does not calculate a reader’s brokerage margin, recommend leverage or derivatives positions, or provide personalized financial advice.