Response to consultation on Reporting for the purposes of the validation and monitoring of ISDA SIMM

Question 1. Do you agree that the criteria set out in Article [1] of this decision support the identification of reporting entities with significant OTC trading activities?

The International Swaps and Derivatives Association, Inc. (ISDA) appreciates the opportunity to respond on behalf of users of the ISDA Standard Initial Margin Model (ISDA SIMM® or SIMM)  to the Consultation Paper issued by the European Banking Authority (EBA) to counterparties regarding proposed reporting requirements for the purpose of validating and monitoring ISDA SIMM (the CP).

In coordination with and for the benefit of derivatives market participants, ISDA developed and maintains the ISDA SIMM to calculate regulatory initial margin (IM) for global regulatory margin requirements for non-cleared derivatives. Market participants who calculate or have their regulatory IM (reg IM) calculated on their behalf using SIMM are required to be licensed with ISDA. Many of these SIMM users have one or more legal entities that are counterparties in the EU and may be subject to the initial margin model authorization and validation requirements (IMMV) as required pursuant to EMIR for which the EBA will act as the central validator.

As ISDA SIMM is currently the only IM calculation methodology that meets the definition of a “pro forma model” under EMIR, SIMM users are the entities which would be required to comply with these reporting requirements.  As SIMM is used broadly as a global model with established practices in various jurisdictions for its authorization and oversight, we urge the EBA to:

• require only the collection of data which is relevant to the performance of SIMM; and
• align the reporting requirements as closely as possible with the testing and reporting SIMM users have been conducting for 10 years for ISDA for industry monitoring of SIMM performance (QM) and directly to regulators for oversight.

5.2.1. The reporting population

Question 1. Do you agree that the criteria set out in Article [1] of this decision support the identification of reporting entities with significant OTC trading activities?

ISDA supports the proportionate use of a two-tiered approach to determining the SIMM users with “significant OTC trading activities” required to report broader and more granular data, which relies on both the group level AANA as well as an entity level materiality threshold.

ISDA supports the group-level criteria for the identification of ‘Reporting entities engaged in significant OTC trading activities’. To ease the burden of recalculating the group-level AANA, where a group has breached the AANA threshold, its entities should be permitted to assume that the threshold continues to be breached on an ongoing basis, unless the firm elects to recalculate its AANA and demonstrates that it no longer exceeds the threshold.

We note that the EBA has yet to calibrate the first prong of the entity level test - an EUR threshold which will be determined based on data currently being collected from SIMM users for the purpose of determining initial fees for SIMM validation. Without knowing what figure will be set as the threshold, we are unable to comment on the extent to which the threshold will reduce reporting burden and promote the capture of meaningful data on SIMM performance. We request that the EBA consult further once the proposed threshold is determined.

We do not support the second prong of the entity level test which would categorize at least one EU entity of a group with an AANA of at least EUR 750 billion as having ‘significant OTC trading activities’ even if that entity fails to meet the established materiality threshold in the first prong.  Doing so may capture EU entities of global groups which engage in limited derivatives trading and are not already conducting extensive testing and reporting with respect to their SIMM use. A substantive new operational burden would disincentivize the group’s use of that entity for trading of derivatives, redirecting activity to non-EU entities.

1. ​The EBA consults on reporting framework for validation and monitoring of ISDA Standard Initial Margin Model | European Banking Authority

Question 2. Do you have suggestions for an alternative criterion, or alternative criteria (including, where relevant threshold(s)), for identifying counterparties with (individually) significant OTC trading activities? If your proposal includes a threshold, please specify both the reference measure (e.g. initial margins, AANA) and the level you deem appropriate (if the threshold is expressed in monetary terms, please specify it in EUR). The population of entities reporting the more comprehensive set of information that is selected based on these criteria should be sufficiently broad and diverse to support identifying weaknesses of the pro forma model, ideally include at least one entity of every group engaged in OTC trading activities in the EU as well as standalone entities with significant OTC trading activities. The criterion or criteria should be easy to assess and monitor, without significant cost of compliance for monitoring the entity’s position towards the threshold. Please explain how your suggested criterion or criteria meet these requirements.

We are strongly of the view that AANA or any notional based criteria would be an impractical as well as inaccurate approach for identifying individual counterparties with “significant OTC trading activities” for the purpose of assigning more frequent and granular SIMM reporting requirements. AANA captures trades not in scope for regulatory IM requirements, and IM calculated via the regulatory schedule is not relevant to the oversight of SIMM.  A criteria based on SIMM amount at legal entity level would be preferred.

ISDA estimates that a threshold of EUR 1bn at legal entity level would capture ~98% of the IM exposure and would cover ~50% of entities that participate in SIMM quarterly monitoring (QM). Here IM has been calculated based on the SIMM-to-Call calculation (without application of thresholds) as reported by firms in the QM July 2026 report.

Question 3. Do you agree with the use of the ‘collateral portfolio indicator’/ UTI as tool to make the link to the EMIR trade repositories reporting? Do you see issues arising when trying to make that link? Are the assumptions regarding the link between netting sets, margined portfolios and these identifiers plausible? Could it happen that a netting set comprises more than one margined portfolio where the initial margin is computed using an initial margin model based on ISDA SIMM, and if yes, are such cases frequent? What other identifiers could be used?

Where a netting set includes transactions margined under a mixture of SIMM and GRID this does not result in two portfolios, but a mixed calculation for a single portfolio.  A netting set is a single margined portfolio.

Collateral portfolio indicator is field 8 in Commission Delegated Regulation (EU) 2022/18 and is only an indication of whether collateralisation is performed at portfolio level, however we believe that the collateral portfolio code, field 9, is intended to be reported as this is the unique portfolio identifier assigned by the entity making the report. We do not anticipate use of the UTI, as it is not possible to exchange regulatory margin a single transaction on a standalone basis without having a full suite of Regulatory IM documentation in place, which would include setting up the margin portfolio internally and assigning an identifier.  As such, we would see the use of UTI as redundant as the portfolio identifier should still be used for reporting regardless of how many transactions are in the portfolio.

Question 4. Do you usually monitor, and keep a record of, the initial margins calculated / collected as of each business day? What is the cost of compliance associated with reporting the maximum and average values for the month?

Data Retention and Monthly vs. Quarterly Calculations: Firms are more likely to retain month-end figures, while daily IM calculations and collection amounts are not retained by parties for a long period. In accordance with current requirements for both the ISDA SIMM Governance Framework for industry monitoring and for reporting required for regulators, testing is conducted and data is reported quarterly based on data from the prior quarter, not data from each month of the prior quarter. Requiring the reporting of monthly data, even on a quarterly frequency, significantly increases data retention and processing requirements for SIMM users. We  request that the EBA requires a single set of data for Activities based on the data from the relevant quarter.

Maximum and Average IM Amounts: ISDA does not support the reporting of monthly maximum and average IM amounts [R: 0170, 0180, 0230, 0240, 0320, 0330] for the following reasons:

• Monthly maximum and average IM amounts are not monitored for the purpose of internal governance of SIMM, which is based on IM amounts as of specific reference dates. Reporting these figures would therefore require firms to build new calculation and reporting processes solely for this purpose.
• We do not believe these figures are relevant to assessing the performance of SIMM, since the appropriateness of each IM amount depends on the relevant portfolio meeting the regulatory minimum.

As an alternative, firms could report the month-end IM amount calculated since these figures are already produced and used for various purposes, including the reporting requirement for fee calculation. 

1. ISDA Publishes Updated ISDA SIMM® Governance Framework – International Swaps and Derivatives Association

Question 5. Do you see a need to provide entities other than those engaged in significant OTC trading activities with the discretion to omit the breakdown by method for calculating the margin in template V 01.00? Are there particular challenges to providing the breakdown of the number of counterparties and the initial margins by method for calculating the margin, i.e. is there a particular cost of compliance associated with providing that breakdown?

Initial Margin Amount Calculated: Entities that do not meet the test for “significant OTC trading activities” should not be required to report the breakdown of their IM amounts based on the calculation method. Entities that trade fewer OTC derivatives are more likely to leverage an asset manager, administrator or vendor to conduct or assist with their IM calculations and may not have ready access to the component calculations or may need to aggregate these based on data from multiple fund managers, greatly increasing their cost and effort to obtain, prepare and report the data.

Initial Margin Amount Collected: The data fields pertaining to the reporting of IM collected broken down by calculation method (R:0210/C:0020-0050) should be eliminated from the Activities template since they cannot be accurately reported by any firms. For each netting set subject to a master netting agreement, usually the ISDA Credit Support Annex or ISDA Credit Support Deed (CSA), each party makes a single daily margin call for IM from their counterparty which includes all IM amounts subject to collection under the CSA (regardless of whether the IM is regulatorily required and without regard to the calculation method) minus the agreed IM threshold.  As a result, the IM collected (or called for) on a given day can be reported in the aggregate but cannot be reported based on the IM calculation method.

For example, an IM call of EUR 50 million could consist of EUR 80 million from the SIMM calculation and EUR 10 million from a regulatory grid (grid) calculation minus an IM threshold of EUR 40 million ((80+10)-40=50).  The daily IM call of EUR 50 million cannot be subdivided because the IM threshold is not allocated based on the IM calculation method.

Another example which is even more challenging is when the CSA allows a party to collect non-regulatory IM for legacy and, potentially, new transactions (referred to as independent amount (IA) or house IM). The IA includes derivatives not subject to reg IM requirements and may be calculated based on a proprietary method. In their reg IM CSAs parties agreed an approach for the relationship between existing IA arrangements and reg IM requirements. If the CSA specifies the “greater of” approach, then the secured party calculates each of the reg IM amount and the house IM to determine their IM call amount. So, to add on to the prior example, if the reg IM calculated (for all calculation methods) was EUR  90 million, the IM collect amount after the IM threshold is EUR 50 million (90-40=50), and the house IM amount was EUR 75 million, the IM call would be EUR 75 million, as the house IA was greater. This amount exceeds the reg IM minimum requirement but still cannot be disaggregated based on the calculation method. 

This is an example of ‘overcollateralisation’ from the perspective of reg IM requirements, but not from a firm’s perspective of abiding by the terms of the CSA.  As differentiating overcollaterisation is beyond the scope of reg IM and not relevant to assessing the use or performance of SIMM, firms should not be required to report the subset of aggregate IM collected which exceeds the IM amount calculated [R:0220].

With respect to the aggregate IM amount collected as of the reference date [R:0210/C:0010], we note that depending on the firm it may be easier to report the total IM called on the reference date (which would be collected on the following business day) rather than the amount collected on the reference date (which was called on the prior business day). As there would not be a substantive difference between these figures, firms should have the discretion to report either to satisfy this field.

Initial Margin Amounts Calculated but Not Collected: This section [R: 0310, 0320, 0330/C: 0010-0050] should be removed from the template.  It is unclear what information these fields are trying to capture, and how it is relevant to an assessment of SIMM. The regulatory IM amount calculated may not be the amount collected for various reasons, including due to the consideration of IA, the IM threshold, and amounts called but not received. 

For the latter, the obligation to manage IM disputes is well established in the portfolio reconciliation requirements in Articles 13 and 15 of Commission Delegated Regulation (EU) No 149/2013, and any such uncollected amounts are not immediately attributable to a particular calculation method. Only material disputes are investigated for a source of discrepancy as part of the resolution process. Where relevant, this information may be reported in templates v 10.00 or v 11.00.

Summary: In summary, the data proposed in R: 0170, 0180, 0220, 0230, 0240, 0310, 0320, 0330/C: 0010-0050, as well as R: 0210/C:0020-0050 should be deleted from the template as they either cannot be reported and/or do not relate to an understanding of the performance of SIMM.

1. Margin-Approaches-New-Edited-August-2022.pdf
2. Commission Delegated Regulation (EU) No 149/2013 of 19 December 2012 supplementing Regulation (EU) 

Question 6. Do you think an (im)materiality threshold should be introduced (i.e. for example, should counterparties only be counted if the total initial margin calculated for all transactions with that counterparty amounts to at least EUR 500 000? Is the determination of the number of counterparties in such a case associated with a high(er) or low(er) operational cost?

For the purpose of the Activity data, it is not valuable to apply a materiality threshold.  Since this information is reported on an aggregate basis, it is easier to sum the relevant figures for all portfolios and provide the corresponding counterparty count of all such portfolios rather than build to exclude a subset.

Question 7: What are your views on the alternative approach of collecting the information on initial margins, rather than the transformed notional amount under the main method and first alternative method? Do you prefer:

Please elaborate:

ISDA supports Option A and does not support Option B. Option B involves transferring much more data and significantly more granular data, requiring greater effort to build and additional internal authorization to release firm-level product class SIMM calculations.

Question 8. Do you consider 500 000 EUR to be an appropriate amount to determine the relevance, and therefore materiality of a formal margin dispute? Would you support a higher/lower threshold, and for which reasons?

A threshold of EUR 500 000 for disputes lasting more than 5 days is not appropriate for reporting margin disputes related to SIMM. We request that the EBA instead set a threshold of EUR 15 million for disputes outstanding for at least 15 business days, consistent with the threshold for reporting margin disputes to the competent authority under the dispute resolution provisions of Article 15 of Commission Delegated Regulation (EU) No 149/2013. This threshold should be applied to disputes outstanding as of each month end. Disputes that have been resolved by the month-end reporting date should not be reportable, even if they met the threshold at some point during the month.

A lower threshold would impose a significant and disproportionate burden. Investigating margin discrepancies is primarily manual and requires significant staffing resources. Although firms’ internal procedures for investigating IM discrepancies vary, only substantive differences are treated as actionable disputes. Moreover, because the secured party calls IM at the netting set level regardless of the calculation method(s), whether a difference is caused by a particular component calculation (e.g., SIMM vs. grid) may only become clear once an investigation is triggered.

ISDA conducted a survey to estimate the number of potential disputes that would be reportable under different thresholds. At EUR 500 000, the industry would report 60% of all portfolios with potential disputes. For most reporting entities, this would mean reporting a significant number of portfolios each quarter, the majority of which will not yet have been investigated for cause or raised with the counterparty for resolution.

We also note that the proposed definition of a ‘formal dispute’ would capture all disputes, whereas the term typically refers to a scenario in which a counterparty states in writing that it is formally disputing a call, which is rare in practice. We propose replacing ‘formal disputes’ with ‘margin disputes’, in line with the terminology used in Commission Delegated Regulation (EU) 2016/2251.

The requirement to include the expected time to resolve is impractical where we have long-standing disputes that arise from an intractable difference in interpretation with our counterparty which we have no expectation of resolving, or where a number of disputes in combination churn and fluctuate. 

Finally, through the existing monthly reporting required under Commission Delegated Regulation (EU) No 149/2013, firms already provide visibility of disputes resolved via the change in monthly reported items. We propose that the EBA rely on this granular monthly data, which will provide more frequent snapshots of live disputes.

Question 9. Considering both the definition of ‘margin disputes’ and the frequency of their occurrence, what are your views on the alternative approach of collecting only granular information, but for all disputes, in V 11.00? From a cost perspective, what is less burdensome for you:

Please elaborate:

ISDA supports the collection of granular information in template V 11.00 only for margin disputes where the disputed portion of the IM amount (i.e. the amount not posted by one of the counterparties) exceeds EUR 15 million and the dispute has been outstanding for at least 15 business days. This aligns with the threshold for reporting disputes to competent authorities under Article 15 of Commission Delegated Regulation (EU) No 149/2013, and reflects the level at which firms actively investigate disputes and would therefore be able to provide the specified data.

Question 10. Considering both the definition of ‘adjustments’ (which may require further clarification) and the frequency of their occurrence, what are your views on the alternative approach of collecting only the granular information, but on all adjustments applied during the reference period?

ISDA recommends that templates V 12.00 and V 13.00 be deleted and replaced with a single revised template that requires only the information firms already submit for remediated portfolios as part of the ISDA SIMM Quarterly Monitoring exercise.

For context, firms report that adjustments to IM calculated using ISDA SIMM are applied primarily to remediate a portfolio bilaterally, where testing evidence demonstrates the need for remediation and that evidence has been validated by the counterparty. As these adjustments are already captured in the information reported for the ISDA SIMM Quarterly Monitoring exercise, aligning the EBA template with that data would give the EBA the information relevant to SIMM performance without creating a duplicative reporting build.

With respect to template V 14.00 (non-use of the model), ISDA supports the template as proposed for reporting instances where the model is not used.

Question 11. For technical reasons, template V 16.00 must have a unique row key. The combination of the ID of the margined portfolio, the horizon type and the portfolio population (static/dynamic) is assumed to be unique. Should you be aware of instances where this combination is not unique (e.g. back-testing performed for one and the same netting set, horizon type and portfolio population, but using different kinds of P&Ls), please describe them in your response.

As background, ISDA allows for the following four types of backtesting in firm submissions to ISDA:

• "A1" - Actual PnL method comparing 1-day PnL with 1-day calibrated SIMM
• "A1S" - Actual PnL method comparing scaled 1-day PnL with 10-day SIMM
• "A10" - Actual PnL method comparing 10-day PnL with SIMM
• "B" - Backtesting 1+3 method in the style of ISDA backtests or similar 

The EBA’s IMMV reporting instructions document states that “the combination of the values in columns 0010, 0110 and 0120 is a row identifier, which must be unique for each row in the template”. Those are different from the stated text in this question, which includes field “(0130) Portfolio population”. We assume that the question is correct and the draft instructions are incorrect in this respect. The portfolio population is required for the key to be unique.

For the four types of backtesting set out by ISDA, their unique IMMV row keys are given in the following table:

ISDA test

type code

Backtesting

Methodology (0110)

Horizon Type

(0120)

Portfolio

Population (0130)

A1

Model developer

1-day

Dynamic

A1S

Model developer

1-day

Dynamic

A10

Model developer

10-day overlapping

Dynamic

B

Model developer

10-day overlapping

Static

We note that the keys are indeed unique, with the exception that A1 and A1S share a key. In practice, we think this is unlikely to be a problem, as it is unlikely that a single firm would submit both A1 and A1S styles of backtest. Typically, a firm runs one style from the 'A'-family, as well as the "B" style. 

We note that the "Call/Post" direction is not included in the key. Should the requirement for V 16.00 require back-testing to be extended to both Call and Post side of a portfolio, then the unique combination on the fields above would not hold. 

However, firms participating in ISDA’s QM submit back-testing on SIMM-to-Call side only. We would request that the scope for the netting sets in V 16.00 is aligned to the same scope applied to portfolios in ISDA’s QM exercise, which further means portfolios in scope are those which are satisfied by Condition 1 and 2 or Condition 1 and 3 defined below:

• Condition 1: the portfolio has been subject to regulatory IM for at least one year, and is therefore more representative of the steady state of portfolios for regulatory IM;
• Condition 2: non-zero collateral has been collected by either party as of the portfolio reference date, which is the last date of the observation period under the actual test and the date the portfolio is observed at under the ISDA Quarterly 1+3 Backtest;
• Condition 3: zero collateral has been collected by both parties as of the portfolio reference date, and the portfolio has a SIMM margin amount of at least EUR 25 million.

Question 12. Do you have any other comments on this proposal for reporting for the purposes of validating and monitoring pro forma models, and authorizing the use of the initial margin models based on them? Are the templates and instructions on how to fill them in clear?

ISDA proposes removing column 0240, Margin Average Shortfall (MAS), from template V 16.00. As set out in ISDA’s response to the Consultation on the IMMA RTS and Guidelines (letter dated 17 June 2026), MAS goes beyond what is required to assess IM model performance against the margin rules. It should be replaced with a metric aligned with the SIMM Green Shortfall under the SIMM Governance Framework, or with an alternative measure suited to assessing model performance against the margin rules. The SIMM Green Shortfall is a more appropriate measure than MAS: it tests SIMM against the regulatory standard and is already used in the ISDA SIMM Quarterly Monitoring exercise. 

ISDA also proposes that the requirement to report 30 netting sets in template V 16.00 be removed and that the template instead adopt the established reporting standard of the ISDA SIMM Quarterly Monitoring exercise. As currently drafted, entities that have, or belong to a group that has, an AANA of at least EUR 750 billion would report a fixed allocation of 15 red-zone, 10 yellow-zone and 5 green-zone netting sets, topped up with yellow-zone or green-zone netting sets where fewer red-zone or yellow-zone netting sets exist. All other entities would report their 30 most relevant margined portfolios, selected either by margin average shortfall or by a materiality criterion of their own choosing. This approach raises two concerns:

• Disproportionate burden: Analyzing 30 netting sets per entity would impose a significant burden on each firm without a corresponding prudential benefit. Because the allocation is topped up with green-zone netting sets where too few red-zone or yellow-zone netting sets exist, an entity whose portfolios all perform within the green zone could still be required to report and analyze 30 netting sets with up to 22 exceedances each, or as many as 660 exceedances in total. Exceedances in green-zone portfolios are consistent with the expected performance of a well-calibrated model and are likely to reflect random noise rather than any weakness in SIMM, so no meaningful conclusions could be drawn about their drivers. Where portfolios fall in the red or yellow zone, the burden would be greater still, as each exceedance would need to be analyzed to identify its underlying driver. In both cases, the workload would be driven by a fixed count of 30 netting sets rather than by the significance of the results.
• Inconsistent selection: Because portfolios would be selected on different bases, two entities with similar portfolios could report entirely different subsets. This would limit the EBA’s ability to compare results across reporting entities or to aggregate them meaningfully when assessing the performance of SIMM.

By contrast, the ISDA SIMM Quarterly Monitoring exercise applies the SIMM Green Shortfall, a measure aligned with assessing whether portfolios meet the regulatory requirement, together with a single, established set of reporting requirements for all participants. Under the SIMM Monitoring Guidelines, participants report each quarter:

1. Red-zone exceptions: all portfolios with a red traffic-light result under the Actual PnL test (one-day or ten-day) or the ISDA Quarterly Backtest, where the shortfall amount is greater than both EUR 10 million and 15% of the SIMM amount for the portfolio;
2. Amber-zone exceptions: all portfolios with an amber traffic-light result under the Actual PnL test (one-day or ten-day) or the ISDA Quarterly Backtest, where the shortfall amount is greater than EUR 100 million; and
3. Top five portfolios: the five portfolios with the worst traffic-light category and highest shortfall amounts under each of the Actual PnL test and the ISDA Quarterly Backtest, ranked first by traffic-light colour (red, then amber, then green) and then by shortfall amount, including where these fall below the thresholds in (a) and (b).

All red-zone and amber-zone exceptions above these thresholds are reported, even where they are not among the top five.

ISDA therefore asks that template V 16.00 adopt these requirements as its reporting standard. Doing so would address both concerns set out above. First, reporting would be proportionate: the number of portfolios reported would be driven by the number of material exceptions rather than by a fixed count of 30, so firms would focus their analysis on red-zone and amber-zone portfolios with significant shortfalls rather than on large numbers of green-zone exceedances that reflect random noise. Second, every reporting entity would select portfolios on the same objective, threshold-based criteria, giving the EBA data that is directly comparable across entities and with the industry-level data already collected by ISDA. Adopting this standard would also allow firms to use the processes and controls they already operate for the ISDA SIMM Quarterly Monitoring exercise, avoiding a parallel selection methodology, and would remove the reliance on margin average shortfall, which, as set out above, should be replaced with a metric aligned with the SIMM Green Shortfall or an alternative measure suited to assessing model performance against the margin rules.

ISDA also proposes that the ‘Value at dispute peak date’ fields be removed from template V 11.00. Firms do not track this metric for disputes, and reporting it would require them to retain and calculate data beyond the quarter-end snapshot of disputes, including for peak dates that may fall outside the reference dates for the quarter.

Upload files

Name of the organization