ISDA SIMM at 10: How a Shared Margin Model Supports Derivatives Markets

ISDA’s Standard Initial Margin Model (SIMM) has become shared infrastructure for derivatives firms: more than 400 groups of entities use it across over 21,000 portfolios, with approximately $1 trillion in initial margin collected through the framework, according to estimates cited by ISDA. A decade after its September 2016 launch, the model’s reach reflects a practical achievement: firms can calculate initial margin using a common method, while retaining a process for monitoring and updating that method as markets change.

In Episode 60 of ISDA’s The Swap podcast, Eduardo Epperlein, Managing Director at Nomura International, and Tomo Kodama, Managing Director at Bank of America Merrill Lynch, discussed why the industry pursued a shared framework, how it has performed through volatility, and where future changes may be needed. ISDA says SIMM is permitted for use in more than 40 countries.

A common method for calculating exposure

Before SIMM, firms faced a choice between a standard margin grid and proprietary models. Kodama said the grid did not adequately account for offsetting risks within portfolios. Firm-built models could produce substantially different margin calculations for the same relationship between counterparties. That combination risked disputes over calls, uneven liquidity demands and added reconciliation work. Smaller firms, in particular, could have struggled to reproduce calculations from larger institutions.

“The alternative world would be much more expensive, more uncertain, less transparent, and more operationally difficult,” Kodama said. Industry participants began raising concerns in 2013 and later asked ISDA to coordinate a common approach. The aim was a model that firms of different sizes could apply and reproduce, with enough sensitivity to reflect portfolio risk. “We always thought that SIM could be used globally by all participants,” Kodama said.

Scott O’Malia, ISDA’s Chief Executive, said coordination and transparency were needed to implement margin rules without market disruption. Building the framework required more than agreeing on a calculation. The work also covered legal documentation, data formats, reconciliation procedures and operational standards. ISDA formed industry committees focused on those areas and on risk data and methodology.

Epperlein said the model adapted and calibrated sensitivity-based methods to measure counterparty exposure. A shared methodology replaced the competing firm-by-firm approaches that participants had initially considered. Its value depends on firms being able to understand the model’s outputs and apply consistent processes around them.

Monitoring the model and its limits

SIMM’s governance includes weekly industry meetings to review performance and consider changes. Epperlein said participants developed a shared backtesting process and a monitoring framework based on data contributed by firms. That approach provides a view across the industry rather than relying only on separate assessments of each institution’s model.

Quarterly backtesting reports are shared with regulators to assess whether the model continues to meet its intended risk coverage. No single mathematical model can capture every possible portfolio without becoming prohibitively complex, Kodama said. SIMM therefore includes a process to identify portfolios that may need extra margin.

In the latest quarterly monitoring exercise cited by Kodama, 13 of more than 11,000 portfolios required remediation through portfolio adjustments or additional collateral to meet the 99% risk coverage standard. The figure offers a snapshot of how often the process identified a shortfall in that exercise; it doesn’t mean every portfolio is free of risk or that the model will cover every future market event.

ISDA has made 17 changes to SIMM over its first decade. O’Malia described that ongoing maintenance as central to keeping the framework relevant. The record points to a continuing trade-off: changes can help the model reflect market developments, while frequent or poorly supported revisions can complicate implementation.

Margin stability during volatile markets

A core design goal has been limiting procyclicality: sharp increases in margin during market stress that can intensify firms’ liquidity needs. Epperlein said the model’s calibration uses at least four years of historical data, with a stressed period accounting for 25% of the sample. The approach is intended to temper reactions to short-lived volatility.

Discussing the March 2020 dash for cash and the UK gilt crisis in 2022, Epperlein said SIMM performed well and required relatively limited adjustments. He also said central counterparty margin models can be more procyclical. “SIM is designed to be less pro-cyclical, a bit conservative, but that’s its strength,” he said.

For firms managing funding and collateral, predictable margin demands can matter as much as the level required in calmer markets. “Sure, it’s great to have low margins in good times, but in bad times, you don’t want those margin levels skyrocketing,” Kodama said. A model that moderates abrupt increases can help with liquidity planning, though it cannot remove the funding demands that market stress brings.

ISDA shifted from annual to semiannual calibration in 2025. O’Malia said the change allows more frequent responses to market conditions on a predictable schedule. Epperlein added that recalibration can now take place as a routine process without separate pre-approval when the underlying methodology stays the same.

New products and infrastructure

Kodama identified additional jurisdictions, especially China, along with new currencies, yield curves and financial products as possible areas for further development. Tokenization and cryptocurrencies could also warrant attention if their use grows. He cautioned that changes need a clear basis: unnecessary adjustments could create inconsistencies in a framework designed to be applied across firms. “We want to be careful as we make changes,” he said.

The service ecosystem has expanded alongside the model. Figures cited by ISDA show that nearly 70 vendors now offer SIMM-related services. Standard risk data and exchange formats have enabled tools for margin reconciliation, portfolio optimization and collateral management. Kodama characterized SIMM as an industry utility, rather than a framework built to benefit one institution.

Epperlein, who began his career as a physicist studying extreme behavior in laser fusion, called SIMM a living model that must respond to changing risks. Kodama recalled its September 2016 launch as an “emotional milestone” and a chance to contribute something constructive to the industry. The next test will be maintaining that shared foundation as markets, products and data evolve, while grounding each change in evidence.