Reader question: A derivative can have almost zero market value today and still create a large counterparty exposure tomorrow. How does the Basel Standardised Approach for Counterparty Credit Risk convert a portfolio of swaps, options, FX trades and collateral agreements into a regulatory Exposure at Default?
The SA-CCR algorithm separates the problem into two pieces:
- Replacement Cost (RC): how much exposure exists if the counterparty defaults now;
- Potential Future Exposure (PFE): a prescribed add-on for how exposure could grow before the position is closed or remargined.
The Basel formula is:
EAD = α × (RC + PFE),
with:
α = 1.4.
The formula is compact. The real work is determining the legally valid netting set, valuing collateral correctly, assigning each trade to risk classes and hedging sets, computing adjusted notionals and supervisory deltas, applying maturity factors, aggregating offsets only where the rule permits them, and then applying the PFE multiplier.
What this page owns — and what it does not
This page owns:
derivative trades + legal netting + collateral + supervisory parameters → RC + PFE → SA-CCR EAD.
It does not replace ISDA SIMM, which estimates initial margin from market-risk sensitivities; CCP default-waterfall algorithms, which allocate losses after a clearing-member default; or collateral optimisation, which chooses what collateral to post.
This is regulatory counterparty-credit mathematics. It is not a credit opinion on any institution and not a recommendation to trade derivatives.
Why mark-to-market alone is not enough
Suppose an interest-rate swap has a current value of zero. If the counterparty defaults at that exact instant and there is no unpaid cash flow, current replacement cost may be close to zero.
But the swap can move substantially in value before its next margin exchange or over its remaining life. SA-CCR therefore adds a forward-looking supervisory component rather than treating zero current value as zero exposure.
This is the basic decomposition:
current exposure + possible future growth.
Step 1: determine the netting set
Basel requires SA-CCR EAD to be calculated separately for each netting set. Trades can offset one another only when the relevant legal netting arrangement is recognised for regulatory capital purposes.
If no enforceable bilateral netting agreement exists, each transaction is effectively treated as its own netting set under the framework.
This creates a hard boundary:
economic hedge ≠ regulatory netting permission.
A simple netting example
Trade A has market value +8.
Trade B has market value −7.
If both sit in one legally recognised netting set:
V = +1.
If they sit under unrelated agreements, the +8 and −7 cannot simply be collapsed into +1 for SA-CCR.
A system that nets by counterparty name rather than legal agreement can materially understate exposure.
Step 2: calculate Replacement Cost for an unmargined netting set
For an unmargined set, the Basel framework constructs replacement cost from the value of derivative transactions and recognised collateral.
In simplified notation:
RC = max(V − C, 0),
where:
- V is the value of transactions in the netting set;
- C is the haircut-adjusted value of net collateral held under the applicable rules.
Replacement cost cannot be negative. If the bank owes the counterparty rather than the reverse, current counterparty exposure is floored at zero.
Step 3: calculate Replacement Cost for margined trades
For a standard margined netting set, Basel recognises that even daily variation margin can leave exposure because of thresholds, minimum transfer amounts and independent collateral terms.
The Basel replacement-cost structure is:
RC = max(V − C, TH + MTA − NICA, 0),
where:
- TH = margin threshold;
- MTA = minimum transfer amount;
- NICA = net independent collateral amount.
The second term prevents the algorithm from assuming perfect zero exposure when the margin agreement itself permits an unsecured amount to remain outstanding.
A Basel-style margin example
Suppose:
- V = 80;
- C = 90;
- TH = 0;
- MTA = 1;
- NICA = 10.
Then:
V − C = −10.
TH + MTA − NICA = −9.
Therefore:
RC = max(−10, −9, 0) = 0.
The Basel Framework publishes examples of this exact logic to show how collateral and margin terms interact.
Step 4: construct the PFE add-on
Potential Future Exposure is:
PFE = multiplier × Aggregate Add-On.
The aggregate add-on is built from trade-level quantities and then aggregated according to asset-class-specific supervisory formulas.
SA-CCR uses five broad asset classes:
- interest rates;
- foreign exchange;
- credit;
- equity;
- commodities.
Diversification benefits across these asset classes are not recognised in the final aggregate add-on; the asset-class add-ons are summed.
Step 5: identify the primary risk driver
Each derivative must be assigned according to its primary risk factor or factors.
An ordinary fixed-for-floating swap belongs to interest-rate risk. A vanilla FX forward belongs to foreign exchange. A single-name credit default swap belongs to credit risk. Equity options belong to equity risk.
Hybrid products can require allocation to more than one asset class if multiple material drivers are present under the standard.
A product-name mapping such as “all swaps = interest rate” is unsafe.
Step 6: calculate adjusted notional
SA-CCR does not use raw notional identically across all products. The framework adjusts trade notional for features such as duration, exchange rates and product structure.
For interest-rate and credit derivatives, supervisory duration converts a contractual notional into an exposure measure reflecting start and end dates. Options additionally use a supervisory delta.
The generic pattern is:
Trade effective notional ≈ adjusted notional × supervisory delta × maturity factor.
The exact formula varies by asset class.
Supervisory delta gives direction and option sensitivity
For linear long/short positions, delta supplies the direction of exposure. For options, SA-CCR uses a supervisory option-delta formula rather than assuming every option contributes full notional.
This matters because a deep out-of-the-money option should not contribute the same effective notional as an at-the-money option of identical face amount.
The supervisory delta is a regulatory approximation, not the bank’s desk Greek.
Step 7: apply maturity factors
The future-exposure horizon differs between margined and unmargined transactions. SA-CCR uses maturity factors to reflect how long exposure can evolve before effective close-out or remargining.
For unmargined transactions, residual maturity is important. For margined portfolios, the margin period of risk becomes central.
Longer effective exposure horizons generally increase the add-on.
Why daily variation margin does not mean zero PFE
Even a daily-margined portfolio can change value between the last margin exchange and close-out after default. Operational delays, disputes and liquidation time prevent exposure from collapsing to zero instantaneously.
SA-CCR therefore retains PFE for margined trades rather than assuming variation margin eliminates all future exposure.
Step 8: aggregate within hedging sets
SA-CCR recognises offsets within defined hedging sets. For interest-rate derivatives, trades are grouped by currency and maturity structure. Other asset classes use their own grouping rules.
The framework deliberately limits cross-trade offset. Two positions that hedge economically can still receive limited recognition if they belong to different supervisory buckets.
That conservatism is a feature of a standardized model rather than an implementation bug.
Interest-rate maturity buckets
The Basel framework groups interest-rate positions into three maturity buckets:
- less than one year;
- one to five years;
- more than five years.
Effective notionals are aggregated within and across buckets using prescribed correlation structure. A five-month swap and a twenty-year swap are therefore not assumed to offset one-for-one simply because they are in the same currency.
Basis and volatility hedging sets receive special treatment
The Basel framework specifies special supervisory-factor treatment for basis and volatility transactions. In the current framework, the relevant supervisory factor is multiplied by one-half for basis-transaction hedging sets and by five for volatility-transaction hedging sets.
This is a useful reminder that product labels alone are insufficient; the economic risk driver changes the prescribed add-on.
Step 9: apply the PFE multiplier
The PFE multiplier recognises that a netting set with excess collateral or a negative current market value has less chance of becoming exposed than an otherwise identical in-the-money set.
The Basel multiplier behaves as follows:
- when V − C ≥ 0, the multiplier is generally 1;
- as V − C becomes more negative, the multiplier declines;
- it never falls below a 5% floor.
One standard mathematical representation is:
multiplier = min(1, Floor + (1 − Floor) × exp[(V − C)/(2 × (1 − Floor) × AddOn)]),
with:
Floor = 0.05.
The floor prevents large excess collateral from driving modeled future exposure all the way to zero.
A multiplier example
Assume Aggregate Add-On = 20 and V − C = −10.
With Floor = 0.05:
Denominator inside the exponent is:
2 × 0.95 × 20 = 38.
Exponent ≈ −10/38 = −0.2632.
exp(−0.2632) ≈ 0.7686.
So:
multiplier ≈ 0.05 + 0.95 × 0.7686 ≈ 0.7802.
PFE ≈ 0.7802 × 20 = 15.60.
Excess collateral reduces PFE, but does not remove it.
Step 10: calculate EAD
Suppose:
- RC = 5;
- PFE = 15.6.
Then:
EAD = 1.4 × (5 + 15.6) = 28.84.
This EAD becomes an input into downstream regulatory capital calculations. SA-CCR itself does not determine the counterparty’s probability of default or loss given default.
SA-CCR and credit risk are connected but not identical
SA-CCR answers:
How large is the regulatory derivative exposure amount?
A credit-risk framework then combines exposure with counterparty credit quality and other parameters to determine capital.
This is why a low-risk sovereign and a risky corporate can have the same SA-CCR EAD for identical derivative portfolios but different final credit-risk capital.
Collateral can reduce RC and PFE differently
Collateral directly affects replacement cost through V − C. It can also lower the PFE multiplier when the netting set is overcollateralized.
But PFE is floored. Therefore:
more collateral does not create unlimited regulatory exposure relief.
Margin terms are algorithmic inputs, not legal footnotes
Threshold, minimum transfer amount and independent collateral can alter RC. Margin period of risk alters PFE.
If a legal document changes but the risk engine retains the old parameters, the model can be numerically perfect and still wrong.
This makes legal-document data part of the computational state.
Inputs and outputs
A robust SA-CCR engine can require:
- trade identifiers and product types;
- counterparty and legal netting-set identifiers;
- trade market values;
- notionals, currencies, start dates and end dates;
- option strike, underlying and volatility inputs needed for supervisory delta;
- margin agreement status;
- threshold and minimum transfer amount;
- variation margin and independent collateral;
- collateral haircuts;
- margin period of risk;
- asset class and hedging-set mapping;
- supervisory factors and correlations;
- regulatory rule version.
Outputs can include RC, aggregate add-on, PFE multiplier, PFE, EAD, asset-class contributions, hedging-set contributions and rule/data exceptions.
Evidence polarity: what supports confidence?
Evidence for a reliable SA-CCR result includes netting sets tied to current legal opinions, trade valuations reconciling to front-office or risk systems, collateral reconciling to margin records, supervisory factors matching the Basel version, maturity buckets and deltas reproducible at trade level, and Basel example portfolios reproduced independently.
Evidence against confidence includes netting by counterparty name only, negative RC, PFE falling below the 5% floor, interest-rate offsets across unrelated currencies, one maturity factor for every trade, option delta fixed at ±1, or collateral balances that differ from the margin system.
Counterexample: a perfect hedge can still generate SA-CCR add-on
Two trades can offset closely in economic P&L but sit in different supervisory hedging sets or maturity buckets. SA-CCR may therefore recognise only partial offset.
This does not prove the hedge is economically bad. It proves the standardized regulatory model has narrower offset rules.
Counterexample: zero RC does not imply zero EAD
A fully collateralized portfolio can have RC = 0 today. PFE remains positive because future market movement during the exposure horizon is still possible.
Therefore:
RC = 0 does not imply EAD = 0.
Counterexample: more collateral cannot drive PFE to zero
The multiplier floor retains at least 5% of the aggregate add-on. Even extreme overcollateralization therefore leaves a nonzero PFE component under the standard formula.
Counterexample: SIMM initial margin and SA-CCR EAD can move differently
SIMM uses sensitivities, risk weights, correlations and concentration rules to estimate initial margin. SA-CCR uses supervisory add-ons and counterparty-exposure mechanics.
A portfolio change can lower SIMM while raising SA-CCR, or vice versa. The two measures solve different problems.
Weak links in implementation
Legal-netting mismatch. Trades are grouped by counterparty instead of enforceable agreement.
Collateral sign error. Posted collateral is treated as collateral held.
Margin-state drift. A netting set changes from margined to unmargined treatment without the engine updating.
Threshold/MTA omission. RC is understated for margined trades.
Wrong risk driver. Hybrid derivatives are forced into one asset class.
Notional-unit error. Millions and units are mixed.
Maturity-bucket boundary error. Trades fall into the wrong interest-rate bucket.
Delta sign error. Long and short option exposure is reversed.
Cross-asset diversification error. Offsets are recognised across asset classes where Basel does not permit them.
Multiplier-floor error. excess collateral drives PFE below 5% of add-on.
Diagnostics: how to test the algorithm
- single-trade test: calculate one unmargined linear derivative by hand.
- netting test: place equal-and-opposite trades inside and outside a recognised netting set.
- RC floor test: create a deeply out-of-the-money portfolio and verify RC never becomes negative.
- margin-term test: vary TH, MTA and NICA and verify margined RC follows the maximum formula.
- multiplier test: push V − C from positive to strongly negative and verify multiplier falls smoothly but not below 5%.
- bucket test: place interest-rate swaps on each maturity boundary.
- option-delta test: compare deep in-the-money, at-the-money and deep out-of-the-money options.
- asset-class test: ensure rates, FX, credit, equity and commodities aggregate without forbidden cross-class diversification.
- Basel replay test: reproduce published CRE99 worked examples.
- independent implementation test: compare a second code path or spreadsheet against the production result for a controlled portfolio.
What would falsify confidence?
Confidence should be withdrawn if EAD cannot be reconstructed from RC and PFE; if netting exceeds legal permission; if PFE falls below the multiplier floor; if trade-level adjusted notionals cannot be traced; if margin terms disagree with legal documentation; or if published Basel worked examples cannot be reproduced.
Alternatives and limits
SA-CCR is a standardized regulatory exposure method. Banks with approval may use the Internal Model Method for relevant transactions, while other contexts use current-exposure, stress or simulation measures for economic counterparty risk.
SA-CCR does not predict the actual future exposure path of a specific portfolio. Supervisory factors and correlations are intentionally standardised. The method trades portfolio-specific realism for comparability, conservatism and implementability.
How this connects to the surrounding knowledge estate
Collateral data connect SA-CCR to collateral optimisation. Margin methodology remains distinct from ISDA SIMM. Cleared trades connect to the CCP default waterfall. Option supervisory delta has a conceptual connection to the option-pricing and delta estate, although SA-CCR uses regulatory rather than desk-calibrated parameters.
Verification and update triggers
Preserve legal-netting opinions, margin-agreement terms, collateral data, trade taxonomy, maturity logic, supervisory parameters and Basel/jurisdictional rule version. Revalidate after legal-document changes, collateral-platform migrations, new derivative products, Basel FAQs, supervisory-factor changes or material breaks between risk and regulatory trade populations.
Primary and high-quality references
- Basel Framework, CRE52 — Standardised approach to counterparty credit risk, current Basel SA-CCR framework.
- Basel Committee on Banking Supervision, The standardised approach for measuring counterparty credit risk exposures, March 2014.
- Basel Framework, Basel Framework full text, including CRE99 worked SA-CCR examples and margin calculations.
- European Banking Authority, Q&A 2022_6354 — SA-CCR exposure value for a netting set subject to a margin agreement, illustrating RC, PFE and multiplier implementation questions.
Educational boundary: This article explains standardized counterparty-exposure mathematics. It does not provide a credit opinion, derivative strategy or personalized financial advice.