SA-CCR: the add-ons · RiskAtoms Academy

By the end you will be able to

1Assign trades to asset classes and hedging sets, and compute each trade's adjusted notional, supervisory duration and maturity factor.
2Compute the supervisory delta of an option and explain what its sign does.
3Aggregate effective notionals: interest rate maturity buckets, the single-factor formula for equity, credit and commodities, and the sum across asset classes.
4Build the add-on of each of RiskAtoms Bank plc's four netting sets, trade by trade.

Why it matters

Corriemoor pays fixed on £250m for ten years and receives fixed on £150m for five. The two swaps point in opposite directions, yet SA-CCR nets them only in part: they sit in different maturity buckets. Saltney's index put covers £400m of the index, but counts for £31.7m once its delta and maturity factor are applied. The add-on is where SA-CCR turns trades into a number, and these choices decide how big it is.

Core explanation

From trades to one add-on

Illustrative dataRiskAtoms Bank plc's capital book

The same made-up UK bank as Level 3: £67bn of assets and £4bn of equity at 31 December 2025, in £ millions. Module 4.1 adds its regulatory layer: own funds, deductions, risk-weighted assets by risk type, buffers, the leverage exposure and MREL, computed from the Level 3 credit, trading, derivatives and operational risk books. Every figure is illustrative, not any bank's data.

Pro forma. The book applies Basel 3.1, the 2017 finalisation of Basel III, to the 31 December 2025 balance sheet. On that date the bank still reported under the old UK CRR. The UK applies Basel 3.1 from 1 January 2027, and the internal model approach for market risk from 1 January 2028 (PRA PS1/26, paras 1.23 and 1.24). So the book shows the bank as it would look under the new rules.

The counterparty data. The four OTC derivative netting sets of the Level 3 derivatives book (M3.4: Counterparty credit risk): Halvergate Moor Manufacturing (HAL, a manufacturer), Corriemoor Bank (COR), Saltney Capital (SAL, a hedge fund) and Ebbsworth Pension Trustees (EBB), all invented. Module 4.1 gave them SA-CCR exposure of £36.9m and counterparty credit risk RWA of £31.2m. That figure is SA-CCR on these four netting sets only, at Basel's α of 1.4.

Asset classes and hedging sets

L4.4.01: SA-CCR: replacement cost and the multiplier took the replacement cost and the multiplier. The add-on is the rest of PFE: a supervisory stand-in for the expected exposure of L3.4.02: Exposure measures. Each trade goes to one of five asset classes by its primary risk driver (CRE52.24, 52.26): interest rates, FX, credit, equity or commodities. Within a class, trades that can offset sit in one hedging set: one per currency for interest rates, one per currency pair for FX, one for all credit and one for all equity trades, and four for commodities (energy, metals, agricultural, other). Across asset classes there is no offset at all (CRE52.25).

One trade: δ × d × MF

Each trade becomes an effective notional, the product of three numbers.

Size: the adjusted notional d. For an interest rate swap it is the notional times a supervisory duration, which discounts each year at 5%: a 10-year swap counts 7.87 years, a 5-year one 4.42, and a 5-year swap starting in a year 4.21. A swap with a week left is floored at ten business days, 0.04 (CRE52.34). For FX it is the foreign leg in pounds; for equity and commodities, units times price.
Direction: the supervisory delta δ. A linear trade is +1 if it gains when its risk factor rises and −1 if it loses: paying fixed is +1, receiving fixed −1. An option's delta comes from a Black-Scholes style formula with a supervisory volatility, so it depends on the normal distribution (M-C06: The normal distribution). Saltney's bought put has δ = −Φ(−d) = −0.2645.
Horizon: the maturity factor. Supervisory factors are calibrated to a one-year horizon. An unmargined trade with less than a year left is exposed for less: MF = √(min(M, 1)), so 0.50 for three months and 1 for a year or more. A margined trade is exposed only for the margin period of risk: 1.5 × √(10 ÷ 250) = 0.3 at the ten-day floor, 0.424 at twenty days (L3.4.06: The margin period of risk).

Netting within a hedging set

Effective notionals are added (M-A08: Summation notation) and the net is multiplied by the asset class's supervisory factor. How much nets depends on the class:

FX: full netting within a currency pair, then 4%.
Interest rates: three maturity buckets per currency. Trades net fully within a bucket; across buckets, the formula uses 1.4 = 2 × 70% for adjacent buckets and 0.6 = 2 × 30% for the outer two. These are correlations (M-C10: Covariance and correlation): a 5-year rate and a 10-year rate mostly move together, a 6-month and a 10-year rate much less. Then 0.5%. A bank may instead ignore offsets across buckets.
Equity, credit and commodities: trades net fully within one entity (one index, one issuer, one commodity type). Entities combine through a single-factor model with a supervisory correlation ρ: a systematic part, where longs and shorts offset, and an idiosyncratic part, which always adds.

Supervisory parameters, CRE52.72 Table 2 (Swiss SIF copy, version effective 1 January 2023)
Asset class	Supervisory factor	Correlation ρ	Option volatility
Interest rates	0.5%	n/a	50%
FX	4%	n/a	15%
Credit, single name (AAA to CCC)	0.38% to 6%	50%	100%
Credit, index (IG / SG)	0.38% / 1.06%	80%	80%
Equity, single name	32%	50%	120%
Equity, index	20%	80%	75%
Commodity: electricity	40%	40%	150%
Commodity: oil and gas, metals, agricultural, other	18%	40%	70%

The correlation cuts both ways. Long £100m of effective notional on one index and short £60m on another give entity add-ons of 20 and −12: an equity add-on of £15.39m, against £32m if they were simply added. Make both long and it rises to £29.18m: a higher ρ helps a hedged book and hurts a one-way one (CRE52.63).

Sort itWhich hedging set?

RiskAtoms Bank plc trades with one counterparty under one netting agreement. Sort each trade into its hedging set.

Pay fixed on a £100m 7-year SONIA swap
Receive fixed on a $50m 2-year SOFR swap
Buy $65m for pounds in one year
A bought put on a UK equity index
A single-stock total return swap on a UK retailer
Long Brent crude futures

The formal version

The add-on (CRE52.24-52.72; the formulas are images on bis.org and were read as text in the March 2014 SA-CCR standard, paras 157, 159, 164 and 166-179; the option delta also in CRR Article 279a):

\begin{aligned} \text{SD}_i &= \max\Big(\frac{e^{-0.05\,S_i} - e^{-0.05\,E_i}}{0.05},\ \tfrac{10}{250}\Big), \qquad d_i^{\text{IR}} = N_i\,\text{SD}_i \\ \delta_i &= \begin{cases} \pm 1 & \text{linear, long or short} \\ \Phi(d) \text{ or } -\Phi(d) & \text{bought or sold call} \\ -\Phi(-d) \text{ or } \Phi(-d) & \text{bought or sold put} \end{cases}, \qquad d = \frac{\ln(P_i/K_i) + \tfrac12\sigma_i^2 T_i}{\sigma_i\sqrt{T_i}} \\ \text{MF}_i &= \sqrt{\frac{\min(M_i,\ 1)}{1}} \ \ (\text{unmargined}), \qquad \tfrac32\sqrt{\frac{\text{MPoR}_i}{250}} \ \ (\text{margined}) \\ D_i &= \delta_i\, d_i\, \text{MF}_i \\ \text{AddOn}^{\text{IR}} &= \sum_{\text{ccy}} 0.5\% \times \sqrt{D_1^2 + D_2^2 + D_3^2 + 1.4\,D_1D_2 + 1.4\,D_2D_3 + 0.6\,D_1D_3} \\ \text{AddOn}^{\text{FX}} &= \sum_{\text{pairs}} 4\% \times \Big|\sum_{i \in \text{pair}} D_i\Big| \\ \text{AddOn}^{\text{EQ}} &= \sqrt{\Big(\sum_k \rho_k\,A_k\Big)^2 + \sum_k \big(1-\rho_k^2\big) A_k^2}, \qquad A_k = \text{SF}_k \sum_{i \in k} D_i \\ \text{AddOn}^{\text{aggregate}} &= \text{AddOn}^{\text{IR}} + \text{AddOn}^{\text{FX}} + \text{AddOn}^{\text{CR}} + \text{AddOn}^{\text{EQ}} + \text{AddOn}^{\text{CO}} \end{aligned}

$S_i, E_i$ : start and end of the period the trade references, in years (S = 0 once it has started); $M_i$ : remaining maturity, floored at ten business days; $N_i$ : notional.
$P_i, K_i, T_i$ : the underlying price, the strike and the latest exercise date; $\sigma_i$ : the supervisory option volatility of Table 2; Φ: the standard normal distribution function.
$D_1, D_2, D_3$ : the sums of $D_i$ over trades ending in under 1 year, 1 to 5 years and over 5 years. Credit uses the equity formula with its own factors; commodities use it within each of the four hedging sets, which are then added.
$A_k$ keeps its sign, so a short entity offsets a long one through the first term.

Try it

Start with Corriemoor's swaps. Move swap B to 10 years: the two now sit in one bucket and net in full. Move it to under a year: only the 30% correlation is left. Then switch to the option view, start from Saltney's put, and slide the strike to see how the delta, and with it the add-on, grows as the put moves into the money.

Try itBuild an add-on

View

Swap A (starts at Corriemoor's T3)

Notional (£m)£250m

Years to maturity10.0y

The bank…

Swap B (starts at Corriemoor's T4)

Notional (£m)£150m

Years to maturity5.0y

The bank…

Netting set

Trade	SD	d (£m)	δ	MF	Bucket	δ·d·MF
Swap A	7.8694	1,967.3	+1	0.300	D3	590.2
Swap B	4.4240	663.6	−1	0.300	D2	−199.1

D1, under 1 year0.00

D2, 1 to 5 years−199.08

D3, over 5 years590.20

Effective notional (1.4 and 0.6 cross terms)472.73

IR add-on = 0.5% × effective notional£2.364m

Without offsets across buckets£3.946m

If all in one bucket£1.956m

Adjacent buckets: they offset with a 70% correlation.

Worked example

The add-ons of RiskAtoms Bank plc's four netting sets

£ millions. Halvergate is unmargined. Corriemoor and Saltney are margined at a ten-day MPoR. Ebbsworth is margined too, but its cap binds (L4.4.01: SA-CCR: replacement cost and the multiplier), so its add-on is the unmargined one.

Trade by trade

Step 1 of 5

1
Halvergate: one swap, one forward
$\begin{aligned} \text{T1: } d &= 100 \times 5.9062 = 590.62, \quad \delta = -1, \quad \text{MF} = 1, \quad D_3 = -590.62 \\ \text{AddOn}^{\text{IR}} &= 0.5\% \times 590.62 = 2.9531 \\ \text{T2: } d &= 65 \div 1.30 = 50, \quad \delta = +1, \quad \text{AddOn}^{\text{FX}} = 4\% \times 50 = 2.00 \\ \text{AddOn} &= 4.9531 \end{aligned}$

Regulatory lens

Basel: CRE52.24-52.72, version effective 1 January 2023, from the March 2014 SA-CCR standard. The UK, EU and US rules use the same pieces: the PRA's Counterparty Credit Risk (CRR) Part and the EU CRR from Article 274 on, and the US agencies' SA-CCR rule. Jurisdictions as at 3 November 2026 as at 3 Nov 2026.

In the EU

CRR Article 279a sets the supervisory delta with the same signs as Basel: +1 for bought calls and sold puts times Φ(type × d), −1 for sold calls and bought puts. The EBA's technical standards cover options on interest rates and commodities where prices can be negative (Article 279a(3)).

Simplified methods

Banks with small derivatives books may use simpler methods instead: the simplified SA-CCR or the original exposure method in the UK and EU (PRA Counterparty Credit Risk (CRR) Part and CRR Articles 273a-273b). The Basel Framework itself has only SA-CCR and the internal model method (L4.4.03: The internal model method and SFTs).

Spark Atoms

Spark AtomWhere 0.5% and 4% come from

Each supervisory factor is calibrated so that the add-on of a single at-the-money linear trade with a one-year maturity matches its effective EPE (CRE52.44). That is why the formula behaves like a one-year exposure, and why the maturity factor scales it down for shorter horizons.

Spark AtomA put that the bank owns but counts as short

Saltney's put has a 26% delta in size but a negative sign: it gains when the index falls, so it is short the index. Had the bank also held long index futures with Saltney, the two would have netted in the same hedging set.

Key takeaways

Trades go to five asset classes and to hedging sets within them; nothing offsets across asset classes (CRE52.25).
Each trade's effective notional is δ × d × MF: adjusted notional (notional × supervisory duration for rates), supervisory delta, and a maturity factor that is √(min(M, 1)) unmargined or 1.5 × √(MPoR ÷ 250) margined.
Interest rates net fully within a maturity bucket and partly across them (1.4 and 0.6). Corriemoor: £2.364m, against £3.946m with no offset.
Equity, credit and commodities combine entities through a single-factor formula with ρ; signed add-ons let hedges offset. Saltney's put: δ −0.2645, add-on £6.347m.
The four add-ons: Halvergate 4.953, Corriemoor 3.564, Saltney 6.878, Ebbsworth 11.804 (£m).

Reaction

Check what stuck

Question 1 · objective 1

What is the supervisory duration of a 10-year interest rate swap that has already started? (Four decimals.)

Question 2 · objective 3

A sterling hedging set has effective notionals D1 = 0, D2 = +100 and D3 = −100 (£m). What is its interest rate add-on, in £m? (Three decimals.)

£m

Question 3 · objective 2

The bank sells $130m forward for pounds in three years at a spot rate of 1.30, in a margined netting set with a 10-day MPoR. What is the FX add-on, in £m?

£m

Question 4 · objective 2

Saltney's bought put has a supervisory delta of −0.26. What does the sign do?

AIt makes the put offset long index positions in the same hedging setBNothing: only the absolute value mattersCIt makes the add-on negativeDIt doubles the maturity factor

Question 5 · objective 4

Why do the asset class add-ons simply add up?

ASA-CCR recognises no diversification between asset classesBAsset classes are assumed perfectly negatively correlatedCBecause α = 1.4 already includes the diversificationDThey are combined with a correlation of 50%

Glossary

Hedging set: Hedging set: the trades within one asset class that may offset each other: one per currency for interest rates, one per currency pair for FX, one each for credit and equity, and four for commodities (CRE52.45).
Adjusted notional (d): Adjusted notional d: the size of a trade in SA-CCR. Notional × supervisory duration for interest rate and credit trades, the foreign leg in the home currency for FX, units × price for equity and commodities (CRE52.33-52.36).
Supervisory duration (SD): Supervisory duration SD: (e^(−0.05S) − e^(−0.05E)) ÷ 0.05 for a trade running from S to E years, floored at ten business days (CRE52.34).
Supervisory delta (δ): Supervisory delta δ: +1 for a long linear trade, −1 for a short one, and for an option a delta from a simplified Black-Scholes formula with a supervisory volatility (CRE52.38-52.42).
Maturity factor (MF): Maturity factor MF: √(min(M, 1 year) ÷ 1 year) for an unmargined trade, with M floored at ten business days; 1.5 × √(MPoR ÷ 250 days) for a margined one (CRE52.48-52.53).
Effective notional (D): Effective notional D: δ × d × MF for a trade, summed within a hedging set or maturity bucket (CRE52.30).
Supervisory factor (SF): Supervisory factor SF: the asset class's calibrated one-year move, from 0.5% for interest rates to 40% for electricity (CRE52.72, Table 2).
Supervisory correlation (ρ): Supervisory correlation ρ: the weight on the systematic factor that lets entities within a credit, equity or commodity hedging set partly offset (CRE52.54).
Maturity buckets: Maturity buckets: under 1 year, 1 to 5 years and over 5 years, by end date, for interest rate trades (CRE52.56).

Deep Dives & Origins

Deep Dives

Basel Committee on Banking Supervision, The standardised approach for measuring counterparty credit risk exposures, March 2014, Annex 4a: five worked examples, including a swaption and a commodity netting set. Intermediate.
L3.4.03: Exposure profiles of common trades: the exposure profiles that the maturity factor and the supervisory duration imitate. Foundation.

Origins

Basel Committee on Banking Supervision, CRE52: Standardised approach to counterparty credit risk, paras 52.24-52.72 and Table 2, version effective 1 January 2023, bis.org and Swiss SIF copy.
Basel Committee on Banking Supervision, The standardised approach for measuring counterparty credit risk exposures, March 2014, paras 157, 159, 164 and 166-179.
Regulation (EU) No 575/2013 (CRR), Article 279a, as amended by Regulation (EU) 2024/1623.
Prudential Regulation Authority, PRA Rulebook: Counterparty Credit Risk (CRR) Part, Articles 271-276, version from 1 January 2027.

Watch this lesson

Video version

Prefer to watch? The same lesson as a 12-minute video will sit here, made from this page's script.