The Standardized Approach for Counterparty Credit Risk (SA-CCR)

The SA-CCR formulas come up when a bank calculates the exposure at default (EAD) of derivatives and reflects it in its capital ratio. They also come up when it checks those figures. The central task is to check how the inputs (market value, collateral, notional, and remaining maturity) map to the output, the EAD.

What this regulation calculates

SA-CCR (the standardized approach for counterparty credit risk) is a method for estimating, inside capital adequacy regulation, the exposure if a derivative counterparty defaults. The result is the exposure at default, or EAD, which the Japanese notice calls the "credit equivalent amount." On March 31, 2014, the Basel Committee on Banking Supervision (BCBS) published the final rule text "The standardised approach for measuring counterparty credit risk exposures"¹. In Japan, the amendment to the notice promulgated on March 23, 2018 has applied since March 31, 2018².

A bank calculates the EAD of derivatives with SA-CCR or with the internal model method (the "expected exposure method" in the notice). Domestic-standard banks that meet the conditions may also use the current exposure method (CEM)³. Overview of the Regulatory Treatment of Counterparty Risk covers where this fits in the regulation as a whole.

The provisions of the notice (the FSA's capital adequacy notice, called "the Notice" below) that set out SA-CCR are as follows.

The Notice's terms correspond to the English terms of the US rule (12 CFR 217.132) as follows⁴.

The EAD formula: replacement cost and PFE

The Notice calculates EAD for each netting set with the following formula⁵.

$$EAD = 1.4 \times (RC + PFE)$$

$RC$ is the replacement cost, and $PFE$ is the potential future exposure (the "potential future exposure amount" in the Notice). A netting set is a group of trades under a legally effective bilateral netting agreement. Without such an agreement, one trade is one netting set. From the EAD, the bank deducts the amount of CVA recognized in its financial accounting (floored at zero)⁶.

Replacement cost

When there is an agreement to exchange collateral (a "margin agreement" in the Notice), the replacement cost is given by the following formula⁷.

$$RC = \max\{V - C,\ TH + MTA - NICA,\ 0\}$$

$V$ is the total market value of the trades in the netting set. $C$ is the collateral received minus the collateral posted, after adjustment by the haircut (the volatility adjustment). However, posted collateral that a third party holds in a segregated account, arranged so that no loss arises in the counterparty's insolvency proceedings, is not counted as posted.

$TH$ is the threshold below which no variation margin is called (the Notice's term is literally "credit limit"), and $MTA$ is the minimum transfer amount (the "minimum delivery collateral amount" in the Notice). $NICA$ is $C$ excluding variation margin. Because of the threshold and minimum transfer amount terms, even when collateral covers the market value ($V - C = 0$), the replacement cost remains up to $TH + MTA - NICA$.

Without a margin agreement, the formula is $RC = \max\{V - C,\ 0\}$.

PFE and the multiplier

PFE is the product of the multiplier and the aggregate add-on⁸.

$$PFE = \text{multiplier} \times AddOn_{\text{aggregate}}$$ $$\text{multiplier} = \min\left\{1,\ 0.05 + (1 - 0.05)\exp\left(\frac{V - C}{2 \times (1 - 0.05) \times AddOn_{\text{aggregate}}}\right)\right\}$$ $$AddOn_{\text{aggregate}} = AddOn_{\text{IR}} + AddOn_{\text{FX}} + AddOn_{\text{Credit}} + AddOn_{\text{Equity}} + AddOn_{\text{Com}}$$

$AddOn_{\text{aggregate}}$ is the sum of the add-ons of the five asset classes: interest rate, foreign exchange, credit, equity, and commodity. They are not offset across asset classes. The multiplier is 1 when $V - C$ is zero or more. When market value is below collateral (when market value is negative, or when collateral exceeds market value), the multiplier falls exponentially, with a floor of 0.05. The multiplier for each value of $V - C$ divided by the aggregate add-on is as follows (calculated from the formula and rounded to two decimal places; this is not a table in the Notice).

The multiplier floor, the alpha of 1.4, and CEM

The multiplier reflects both the amount by which collateral exceeds market value and the amount by which market value is negative. In the final rule published on January 24, 2020, US authorities explain the floor of 0.05 as keeping PFE from falling to zero however much collateral there is, and as intended to leave some capital even when collateral exceeds market value⁹.

In the same final rule, the authorities explain the alpha of 1.4 as providing for risks that SA-CCR does not fully capture, so that SA-CCR values generally do not fall below internal model method values¹⁰.

In the US, the current exposure method (CEM) was introduced in 1989 and last substantially updated in 1995. The authorities explain implementing SA-CCR in the US as making the measure more risk-sensitive and better calibrated than CEM, and as responding to concerns that CEM has not kept up with the market practice of large banks¹⁰.

How an add-on is calculated

An add-on is calculated by finding the effective notional of each trade and aggregating it by hedging set and asset class. There are four steps¹¹.

1. Assign each trade to interest rate, foreign exchange, credit, equity, or commodity derivatives according to the main factor that moves its market value (the risk driver). A trade with several risk drivers can be assigned to several asset classes.
2. Within an asset class, divide the trades into hedging sets. For interest rates, the hedging set is each currency; for foreign exchange, each currency pair; for commodities, each of energy, metals, agricultural products, and other. Credit and equity each form one hedging set. Trades that reference a basis (the difference between different risk factors in the same currency) form a separate hedging set for each combination of the same risk factors, and trades that reference volatility form a separate hedging set for each asset class.
3. Find the effective notional of each trade.
4. Multiply the effective notional by a supervisory factor, aggregate within the hedging set, and add up within the asset class.

Figure description

The SA-CCR calculation moves back and forth among four units (the trade, the hedging set, the asset class, and the netting set) and ends by combining everything into the EAD of the whole netting set. The third step in the figure (find effective notional) returns to the trade column because the effective notional is a value for each trade. The final step, 1.4 × (RC + PFE), uses the replacement cost found in the chapter on the EAD formula. The effective notional and the supervisory factor are defined in the formulas below.

The effective notional is the following product.

$$\text{effective notional} = \text{adjusted notional} \times \text{delta adjustment} \times \text{maturity adjustment}$$

The adjusted notional is the notional amount multiplied by the duration adjustment for interest rates and credit; the number of units of the underlying multiplied by its price on the calculation reference date (the mark-to-market adjusted notional) for equity and commodities; and the notional amount for foreign exchange¹².

The duration adjustment is given by the following formula, where $S$ is the number of years until the start of the period the trade references and $E$ is the number of years until its end (both are business days converted to years; $S$ is 0 if the start has passed)¹³.

$$\text{duration adjustment} = \frac{\exp(-0.05 S) - \exp(-0.05 E)}{0.05}$$

If this value is below the value of 10 business days converted to years, the latter is used.

The delta adjustment is +1 or −1 for a trade that is neither an option nor a synthetic securitization in credit. It is +1 for a trade whose market value rises when the referenced factor (an interest rate, foreign exchange rate, share price, commodity price, and so on) rises, and −1 for a trade whose market value falls. In credit, "rises" corresponds to a deterioration in the creditworthiness of the reference entity. Synthetic securitizations use a separate formula for each tranche, and options use the formula in the deep dive below¹⁴.

For a trade without a margin agreement, the maturity adjustment is calculated from the remaining maturity $M$ (in business days) with the following formula. For a trade with a margin agreement, it is calculated from the margin period of risk (MPOR)¹⁵.

$$\text{maturity adjustment} = \sqrt{\frac{\min\{M,\ 250\}}{250}} \quad (\text{no margin agreement})$$ $$\text{maturity adjustment} = \frac{3}{2}\sqrt{\frac{MPOR}{250}} \quad (\text{with a margin agreement})$$

Without a margin agreement, the value is capped at 1 once the remaining maturity is one year (250 business days) or more. With a margin agreement, the value is 0.30 when the MPOR is 10 business days. For trades with a long remaining maturity, this difference is why variation margin makes the add-on smaller. From the formulas, the margin agreement gives the smaller value only when $\min\{M,\ 250\}$ exceeds 2.25 times the MPOR. For trades with a short remaining maturity, it is the other way round. For example, with a remaining maturity of 20 business days and an MPOR of 10 business days, the value is $\sqrt{20/250} = 0.28$ without a margin agreement and 0.30 with one. How the MPOR is set is covered below in "Netting and collateral."

Deep diveThe delta adjustment formula for options

The delta adjustment of an option is calculated from the following $d$ and $d'$¹⁴.

$$d = \frac{\ln(P/K) + 0.5\,\sigma^2 T}{\sigma\sqrt{T}}, \qquad d' = -d = \frac{-\ln(P/K) - 0.5\,\sigma^2 T}{\sigma\sqrt{T}}$$

$P$ is the level the option references, $K$ is the strike, $T$ is the number of business days until the latest exercise date converted to years, and $\sigma$ is the volatility the Notice sets for each asset class. $\Phi$ is the cumulative distribution function of the standard normal distribution.

Trade	Delta adjustment
Long call option	$\Phi(d)$
Short call option	$-\Phi(d)$
Long put option	$-\Phi(d')$
Short put option	$\Phi(d')$

Here a call is an option whose market value rises when the referenced level rises, and a put is an option whose market value falls when the level rises. For credit options, read the level as the creditworthiness of the reference entity (a corporate or an index), and "rises" as "deteriorates."

Add-ons by asset class

The supervisory factor, the option volatility, the correlation, and how to aggregate within a hedging set are set for each asset class¹⁶.

For corporates and similar entities, the credit supervisory factor has six levels, from 0.38 to 6.00, matching credit risk categories 1-1 to 1-6, which correspond to the rating of the reference entity. For indices, it is 0.38 if the category is 4-3 or 5-3 or better, and 1.06 if it is below. The supervisory factor for a hedging set that references a basis is half the usual one for interest rates, equity, and commodities. The supervisory factor for a hedging set that references volatility is five times the usual one for interest rates, foreign exchange, equity, and commodities (for interest rates, 2.50 against 0.50).

The aggregation methods are as follows, in the order of foreign exchange, credit and equity, commodities, and interest rates. Credit, equity, and commodities use the same form of formula, which uses a correlation.

For foreign exchange, the effective notionals within a hedging set (a currency pair) are added with the direction of the exchange rate aligned, and 4% is applied to the absolute value.

For credit and equity, the whole netting set is one hedging set. For each reference name (a corporate or an index in credit; a stock or an index in equity), find the add-on $AddOn_k$ (the sum of effective notionals multiplied by the supervisory factor), and combine them with the correlation $\rho_k$ in the following formula.

$$AddOn = \sqrt{\left(\sum_k \rho_k \times AddOn_k\right)^2 + \sum_k \left(1 - \rho_k^2\right) \times AddOn_k^2}$$

If the correlation is 1, the result is the simple sum of the add-ons of the reference names, and if it is 0, the result is the square root of the sum of squares.

For commodities, for each hedging set (energy, metals, agricultural products, other), the add-ons of commodity type $k$ are combined with the same form of formula using a correlation of 0.4, and the add-ons of the hedging sets are added.

For interest rates, the trades within a hedging set (a currency) are divided into three buckets by remaining maturity. Let $D_1$, $D_2$, and $D_3$ be the sums of effective notionals of trades whose referenced interest rate contract ends in less than one year, from one year to five years, and more than five years. They are combined with either formula 1 or formula 2 below, and 0.50% is applied. Currencies are added, not offset. As of October 2026, the Notice's provision on interest rate aggregation lists formula 1 and formula 2 side by side and says only that either may be used. It does not state a condition for choosing between them¹⁷.

$$\text{formula 1}: \sqrt{D_1^2 + D_2^2 + D_3^2 + 1.4\,D_1 D_2 + 1.4\,D_2 D_3 + 0.6\,D_1 D_3}$$ $$\text{formula 2}: |D_1| + |D_2| + |D_3|$$

Deep diveOffsetting within an interest rate hedging set (formula 1 and formula 2)

Consider a hypothetical example in which a netting set has two yen interest rate swaps. The first has a notional amount of 10 billion yen and 5 years remaining, and is on the side whose market value rises when interest rates rise (delta adjustment +1). The second has a notional amount of 10 billion yen and 10 years remaining, and is on the side whose market value falls when interest rates rise (delta adjustment −1). Both have already started, and there is no margin agreement.

The duration adjustment of the first is $(1 - e^{-0.25}) / 0.05 = 4.424$, so its effective notional is 44.24 billion yen (delta +1). Its 5 years remaining fall in $D_2$. For the second it is $(1 - e^{-0.5}) / 0.05 = 7.869$, so its effective notional is −78.69 billion yen (delta −1). Its 10 years remaining fall in $D_3$. The maturity adjustment is 1 for both. $D_1 = 0$.

Formula 1 gives $\sqrt{44.24^2 + 78.69^2 - 1.4 \times 44.24 \times 78.69} = 57.24$ and formula 2 gives $44.24 + 78.69 = 122.93$ (both in billions of yen). Applying 0.50% gives add-ons of 286 million yen under formula 1 and 615 million yen under formula 2. Formula 1 allows trades in opposite directions to offset between adjacent maturity buckets. Formula 2 does not offset between buckets and adds the absolute values.

Numerical example: S Bank's interest rate swap

Consider a hypothetical example in which S Bank has a single yen interest rate swap in a netting set with a counterparty. The notional amount is 10 billion yen, the remaining term is 5 years, and the swap is assumed to be on the side whose market value rises when interest rates rise. The trade has already started, so $S = 0$ and $E = 5$. Taking 250 business days as one year, 5 years remaining is 1,250 business days.

Because the example assumes the side whose market value rises when interest rates rise, the delta adjustment is +1. The duration adjustment is $(1 - e^{-0.25}) / 0.05 = 4.424$, so the adjusted notional is $10 \times 4.424 = 44.24$ billion yen. The trade falls in the yen interest rate hedging set, in the bucket for remaining maturity of one year to five years.

The example compares three cases. In A, there is no margin agreement and the market value is +100 million yen. In B, there is no margin agreement and the market value is −300 million yen. In C, there is an agreement to mark to market daily and exchange variation margin, the market value is +100 million yen, and S Bank has received 100 million yen of cash variation margin. In C, the threshold is 0, the minimum transfer amount is 10 million yen, there is no initial margin, and the margin period of risk is 10 business days.

In A, the market value is positive and there is no collateral, so the replacement cost is 100 million yen and the multiplier is 1. In B, the market value is negative, so the replacement cost is 0, and the multiplier falls to 0.52, so PFE is also smaller. In C, variation margin makes $V - C$ zero, and the replacement cost falls to $\max\{0,\ 0 + 10 - 0,\ 0\} = 10$ million yen. With the maturity adjustment $\frac{3}{2}\sqrt{10/250} = 0.30$, the add-on is 0.30 times that of A, at 66 million yen¹⁸.

Even in B, where S Bank's market value is negative (the side that pays the counterparty), the add-on remains, so EAD does not become zero. The more negative the market value, the closer the multiplier gets to 0.05 and the smaller PFE becomes, but there is a floor.

Netting and collateral

Netting works at two levels. Within a netting set, the market value $V$ is offset between trades. Within a hedging set, effective notionals are offset against trades in the opposite direction. In the interest rate hedging set, however, the maturity buckets are separate, and there is no offset between currencies or between asset classes. A trade without a legally effective bilateral netting agreement is a netting set on its own, and market values are not offset⁶.

Collateral enters in three places: $C$ in the replacement cost formula, the multiplier, and the maturity adjustment. Variation margin narrows the gap between market value and collateral and lowers the replacement cost. For a trade with a margin agreement, the maturity adjustment is determined by the MPOR instead of the remaining maturity. The MPOR (margin period of risk) is the period until the market risk hedge that follows a counterparty default is complete. The number of business days for each category of netting set is as follows¹⁹.

$F$ is the number of business days determined by the four categories above. If there have been disputes, this number of business days changes¹⁹. The condition is that, in the two most recent consecutive quarters before the quarter containing the calculation reference date, there were three or more cases in which a dispute over the adjustment of collateral caused the close-out period to take longer than the number of business days above.

If the condition is met, then for the next two consecutive quarters the MPOR of that netting set is at least twice the number of business days above¹⁹.

The MPOR also enters the volatility adjustment of the collateral itself. With a margin agreement, the adjustment is the standard 10-business-day adjustment multiplied by $\sqrt{MPOR / 10}$²⁰.

Consider a hypothetical example in which Q Bank's netting set with a counterparty is in each of the five states below, to see how collateral moves the replacement cost. All collateral is cash in the same currency as the trades, and the volatility adjustment is 0%²¹. $NICA$ is defined in the Notice only as a symbol: $C$ excluding variation margin⁷.

State 1 has no collateral agreement, so the replacement cost is $\max\{V - C, 0\}$, or 1,000 million yen (1 billion yen). In state 2, the variation margin Q Bank has received covers most of the market value, and the replacement cost falls to 50 million yen. State 3 has a threshold of 500 million yen, so part of the value stays uncovered by collateral, and the replacement cost is 550 million yen. In state 4, the market value from Q Bank's side is −300 million yen, but Q Bank has posted 200 million yen of initial margin that is not segregated, so $NICA$ is −200 million yen, $TH + MTA - NICA$ is 250 million yen, and the replacement cost is 250 million yen. In state 5, where the posted initial margin is segregated by a third party, the replacement cost stays at 50 million yen. The 200 million yen difference between states 4 and 5 is the amount of the non-segregated initial margin.

When several netting sets are under one margin agreement, they are taken together to calculate one EAD. The replacement cost is calculated with a formula that separates the positive and negative market values of each netting set and compares them with the collateral. PFE is the sum of the PFEs of the netting sets, each calculated as if there were no margin agreement²².

Application in Japan and the US

The Notice adopted SA-CCR four years after the BCBS published its document.

In the Notice as of October 2026, only domestic-standard banks that meet the conditions may use CEM. A bank that used SA-CCR on the most recent calculation reference date continues to use SA-CCR, except where it has notified the authority that it cannot continue for unavoidable reasons³.

The EAD from SA-CCR also feeds other regulations. In the CVA risk basic approach (BA-CVA), a bank that uses SA-CCR uses, for the EAD of each netting set, the amount before deducting the effect of CVA²⁴. The calculation is covered in The CVA Risk Capital Charge. In the BCBS final document of December 2017, the leverage ratio came to measure derivative exposure with a modified version of SA-CCR²⁵.

In the US, the Federal Reserve Board's rule 12 CFR 217.132(c) sets out SA-CCR. The skeleton of the formulas is the same as in the Notice: the exposure amount is 1.4 × (replacement cost + PFE), and PFE is the product of the multiplier and the aggregate add-on. The supervisory factors for interest rates, foreign exchange, equity, and commodities also match the Notice's table: 0.50% for interest rates, 4% for foreign exchange, 32% for single stocks, 20% for stock indices, 40% for electricity, and 18% for other commodities²⁶. The US rule has the following provisions.

According to the US authorities' explanation, they removed the alpha for commercial end-users because commenters raised a concern that commercial end-user trades are directional and often uncollateralized, so the exposure would increase compared with CEM. The authorities pointed out that trades with commercial end-users have credit risk mitigants that are not recognized as financial collateral, and that the direction of the risk is likely to work in the bank's favor. They stated that removing the alpha improves the calibration of SA-CCR.

The authorities capped the amount for netting sets with a variation margin agreement for the following reason. For example, for trades with a short remaining maturity of 10 business days or less, the maturity adjustment on the collateralized side is 1.5 times larger. In some cases the amount is then larger than without collateral. The authorities consider such cases rare and the effect on capital small³¹.

The US rule took effect on April 1, 2020 under the final rule published on January 24, 2020. For banking organizations that use the advanced approaches, the mandatory compliance date was January 1, 2022³². The application dates differ from those of the Notice, since each jurisdiction sets its own.

Practical implications

The inputs for each netting set are the market value of each trade, the notional amount, the start and end dates, the delta, the asset class and hedging set, and the category for the supervisory factor. A netting set with a margin agreement also needs the threshold, the minimum transfer amount, the type and currency of the collateral, whether it is segregated, and the frequency of marking to market⁷. The MPOR category needs the number of trades and the history of disputes¹⁹.

What this page does not cover

This page does not cover the detailed differences between the CEM and SA-CCR formulas, or whether the US rule's provisions for commercial end-users, for netting sets with a variation margin agreement, and for the MPOR floor of client-facing trades also appear in the Notice.

Notes

1. Financial Services Agency, "バーゼル銀行監督委員会による最終規則文書「カウンターパーティ信用リスクエクスポージャーの計測に係る標準的手法」の公表について" (On the publication by the Basel Committee on Banking Supervision of the final rule text "The standardised approach for measuring counterparty credit risk exposures", our translation) (April 2, 2014) (in Japanese). It states that the BCBS published the final rule text on March 31. ↩ ↩²

2. Financial Services Agency, "「自己資本比率規制（第１の柱・第３の柱）に関する告示の一部改正（案）」等のパブリックコメントの結果等について" (On the results of the public comment on the draft partial amendments to the notices on capital adequacy regulation (Pillar 1 and Pillar 3), and related matters, our translation) (March 23, 2018), section 3, "公布・適用日" (Promulgation and application dates) (in Japanese). ↩ ↩² ↩³

3. Financial Services Agency, "銀行法第十四条の二の規定に基づき、銀行がその保有する資産等に照らし自己資本の充実の状況が適当であるかどうかを判断するための基準" (Standards for determining whether a bank's capital adequacy is appropriate in light of the assets it holds, under Article 14-2 of the Banking Act, our translation; FSA Notice No. 19 of 2006; the FSA's full version) (in Japanese), Article 79, paragraphs 1 to 3 (pages 188–189). This page calls this the Notice, and the same Notice's provisions below are cited by the printed page numbers. ↩ ↩²

4. The Notice, Article 79-2, paragraph 11, item 2 (supervisory factor), item 5 (duration adjustment), item 6 (delta adjustment), and item 7 (maturity adjustment). The US terms are in 12 CFR 217.132(c)(9)(i) to (iv) and Table 3. ↩

5. The Notice, Article 79-2, paragraph 1 (page 190). When one netting set has several margin agreements, the trades covered by each agreement are grouped together, and the EAD is calculated for each group. ↩

6. The Notice, Article 79, paragraph 7 (pages 189–190). ↩ ↩²

7. The Notice, Article 79-2, paragraph 2 (pages 190–191). $C = C_{\text{collect}} \times (1 - Hc_{\text{collect}} - Hfx_{\text{collect}}) - C_{\text{post}} \times (1 + Hc_{\text{post}} + Hfx_{\text{post}})$. $C_{\text{collect}}$ is the eligible financial collateral received, $C_{\text{post}}$ is the collateral posted (excluding collateral segregated by a third party), $Hc$ is the volatility adjustment for the type of collateral, and $Hfx$ is the adjustment when the exposure and collateral are in different currencies. $NICA$ is defined as "the same as C in the preceding item, but excluding variation margin" (our translation). ↩ ↩² ↩³

8. The Notice, Article 79-2, paragraph 6 (pages 192–193). In the multiplier formula, the Notice's exponential notation is rewritten as $\exp$ for the exponential function. ↩

9. Federal Register, "Standardized Approach for Calculating the Exposure Amount of Derivative Contracts" (final rule of January 24, 2020; "the final rule" below), I.B (the purpose of the multiplier), and III.D, item 3, "PFE Multiplier" (the purpose of the floor). ↩

10. The final rule, SUMMARY; I.C, "Overview of the Proposal" (the year CEM was introduced and last updated, and the position of SA-CCR); and the alpha factor box in II, "Overview of the Final Rule" (the purpose of the alpha). ↩ ↩²

11. The Notice, Article 79-2, paragraphs 7 to 10 (pages 193–194). Paragraph 7 says that a trade with several risk drivers can be assigned to several asset classes at the same time. ↩

12. The Notice, Article 79-2, paragraph 11, items 4 and 5 (interest rate, pages 194–195), paragraph 12, item 3 (foreign exchange, page 197), paragraph 13, item 4 (credit, page 199), paragraph 14, item 4 (equity, page 202), and paragraph 15, item 3 (commodity, page 204). ↩

13. The Notice, Article 79-2, paragraph 11, item 5 (pages 194–195). Credit applies it by paragraph 13, item 5 (page 199). ↩

14. The Notice, Article 79-2, paragraph 11, item 6 (pages 195–196). Foreign exchange applies it by paragraph 12, item 4, credit by paragraph 13, item 6, equity by paragraph 14, item 5, and commodity by paragraph 15, item 4, with the volatility read accordingly. ↩ ↩²

15. The Notice, Article 79-2, paragraph 11, item 7 (page 196). For a trade whose underlying is delivered as a derivative, the remaining maturity $M$ is the number of business days from the maturity date of that underlying to the calculation reference date (10 business days if that number is less than 10). ↩

16. The Notice, Article 79-2, paragraph 11, items 2 and 3 (interest rate), paragraph 12, items 1 and 2 (foreign exchange), paragraph 13, items 1 to 3 (credit), paragraph 14, items 1 to 3 (equity), paragraph 15, items 1 and 2 (commodity), and paragraphs 8 to 10 (hedging sets). ↩

17. The Notice, Article 79-2, paragraph 11, item 3 (page 194), says "calculated using either the formula in (a) or (b)" (our translation). ↩

18. The calculation is done without rounding, and the table shows values to the nearest million yen. The EAD of A is $1.4 \times (100 + 221.2) = 449.68$ million yen, of B is $1.4 \times 113.98 = 159.57$ million yen, and of C is $1.4 \times (10 + 66.36) = 106.90$ million yen. ↩

19. The Notice, Article 79-2, paragraphs 4 and 5 (page 192). ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶

20. The Notice, Article 79-2, paragraph 3, item 2 (page 191). The adjustment without a margin agreement is set by a different formula in item 1 of the same paragraph (page 191). ↩

21. The Notice, Article 94, the table row "現金及び自行預金" (cash and deposits with the bank itself) (page 232). ↩

22. The Notice, Article 79-2, paragraphs 16 to 18 (pages 205–206). ↩

23. Financial Services Agency, "自己資本比率規制（第１の柱・第３の柱）に関する告示の一部改正（案）等の公表について" (Publication of draft partial amendments to the notices on capital adequacy regulation (Pillar 1 and Pillar 3), and related documents, our translation) (December 22, 2017), section 1, "自己資本比率規制に関する告示等" (Notices on capital adequacy regulation, etc.) (in Japanese). ↩

24. The Notice, Article 270-3-3, paragraph 2 (pages 396–397). For $EAD_{NS}$, it provides that when SA-CCR is used, the amount is the credit equivalent amount under Article 79-2, paragraph 1, and the effect of CVA is not taken into account. ↩

25. Financial Services Agency and Bank of Japan, "「レバレッジ比率の枠組みの見直し」の概要" (Overview of "Revisions to the leverage ratio framework", our translation), page 5, "デリバティブ・エクスポージャー（修正 SA-CCR の採用）" (Derivative exposure (adoption of the modified SA-CCR)) (in Japanese). The 2017 final document sets exposure = 1.4 × (RC + PFE) with the modified SA-CCR. ↩

26. 12 CFR 217.132(c)(5)(i) (exposure amount), (c)(7) (PFE), and Table 3 to § 217.132 (supervisory factors). The text was retrieved through the eCFR API on October 5, 2026. ↩

27. 12 CFR 217.132, Table 3 (Credit, single name: Investment grade 0.46, Speculative grade 1.3, Sub-speculative grade 6.0). For the Notice, see Article 79-2, paragraph 13, item 2 (a) (pages 198–199). ↩

28. 12 CFR 217.132(c)(5)(iv). The definition of "commercial end-user" is in another part of the rule. ↩

29. 12 CFR 217.132(c)(5)(ii). Agreements under which the counterparty is not obliged to post variation margin are excluded. ↩

30. 12 CFR 217.132(c)(9)(iv)(A)(2)(ii). For transactions that are not client-facing, the floor is set in (A)(2)(i) as 10 business days plus the re-margining frequency minus 1. ↩

31. The final rule, III.A, "Exposure Amount" (commercial end-users and the alpha), and III.C, "Replacement Cost", footnote 75 (the cap on the amount for netting sets with variation margin). ↩

32. The final rule, DATES and II.B, "Effective Date and Compliance Deadline". The effective date is April 1, 2020, and the mandatory compliance date for advanced approaches banking organizations is January 1, 2022. ↩

33. With an MPOR of 20 business days, the maturity adjustment is $1.5 \times \sqrt{20/250} = 0.4243$, the effective notional is $44{,}239.84 \times 0.4243 = 18{,}769$ million yen, and the add-on is $0.005 \times 18{,}769 = 93.85$ million yen. The replacement cost is 10 million yen and the multiplier is 1 ($V - C = 0$), so the EAD is $1.4 \times (10 + 93.85) = 145.39$ million yen. ↩