What is SIMM and who must use it?

SIMM (Standard Initial Margin Model) is the ISDA-published model for calculating initial margin on uncleared OTC derivatives under the BCBS/IOSCO Uncleared Margin Rules (UMR). It applies to bilateral derivatives between covered entities — banks, broker-dealers, and large funds — that are subject to mandatory margin exchange. Phase 6 (September 2022) brought in firms with aggregate average notional outstanding (AANA) above €8 billion, capturing most active derivatives users. SIMM calculates margin as a risk-weighted sum of delta, vega, and curvature sensitivities, producing a single IM figure that both counterparties agree on, avoiding disputes.

How does SIMM differ from CCP initial margin?

CCP initial margin (such as CME SPAN or LCH SwapClear's PAIRS model) is calculated by the clearinghouse using proprietary scenario-based or VaR models, and covers cleared products only. SIMM is calculated bilaterally by each counterparty for uncleared derivatives — both sides must independently compute and agree on the IM amount. SIMM uses a sensitivity-based approach (delta, vega, curvature) with prescribed risk weights and correlations, while CCP models typically use historical VaR or scenario-based approaches. For equivalent interest rate swap portfolios, SIMM IM is typically 1.2–1.5× CCP initial margin, which is part of what drives clearing economics.

What are the SIMM risk classes?

SIMM has six risk classes: Interest Rate (IR), Credit Qualifying (CRQ), Credit Non-Qualifying (CRNQ), Equity (EQ), Commodity (CM), and FX. Each risk class has its own prescribed risk weights, tenor buckets, and correlation parameters. For a standard interest rate derivatives portfolio, only the IR risk class applies. For a cross-currency swap, both IR and FX apply. For a CDS, the CRQ risk class applies. The total SIMM IM is the sum of IM contributions from each applicable risk class (no cross-class diversification benefit).

How often does ISDA update SIMM parameters?

ISDA recalibrates SIMM annually — risk weights, correlations, and concentration thresholds are reviewed each year using recent market data. The recalibration is published as a new SIMM version (e.g., SIMM 2.6, 2.7) and is typically effective January 1 of the following year, with a transition period. Banks must update their SIMM calculation engines to the new parameters by the effective date. Significant market stress periods (e.g., 2022 rate moves) can cause material recalibrations that increase IM for rate-sensitive portfolios.

What is the IM threshold under UMR and how does it work?

The UMR framework includes a €50M initial margin threshold per counterparty group. This means firms do not need to exchange initial margin until the calculated SIMM IM for a netting set exceeds €50M. Below the threshold, firms must still calculate SIMM but need not post or collect IM. The threshold is calculated at the counterparty group level (consolidated group, not legal entity), and both firms in a bilateral relationship independently apply their own €50M threshold — so both must have calculated IM above their respective thresholds before exchange is required. This creates an asymmetric situation for smaller portfolios.

How does SIMM interact with MVA on the XVA desk?

SIMM IM is the primary input to MVA (Margin Valuation Adjustment), which prices the funding cost of posting initial margin over the life of a trade. When structuring a new uncleared OTC derivative, the XVA desk needs to estimate the expected SIMM IM profile over the trade's life — how IM will evolve as the trade moves in or out of the money, as rates change, and as vega changes for options. This expected IM profile is discounted at the firm's funding spread to produce MVA. Claude can estimate the SIMM IM for specific trade structures, which is the key input to MVA pricing.

SIMM Initial Margin Calculation with Claude AI (UMR 2026)

ISDA SIMM initial margin calculation under UMR: delta, vega, and curvature sensitivities, risk weights, concentration thresholds, cross-currency aggregation, and MVA pricing. Worked Claude AI prompts for IM calculation, UMR compliance, and SIMM dispute resolution.

Educational content, not professional advice — AI output and figures here can be wrong. Verify before you rely on it. Full disclaimer →

SIMM and the UMR Initial Margin Framework

The BCBS/IOSCO Uncleared Margin Rules (UMR) introduced mandatory bilateral initial margin exchange for uncleared OTC derivatives, phased in from 2016 to 2022. ISDA's Standard Initial Margin Model (SIMM) is the industry-standard methodology for these calculations — it is accepted by regulators across all major UMR jurisdictions (EU, UK, US, Japan, Canada, Australia) and used by the vast majority of covered entities. Understanding how SIMM works is essential for any derivatives professional dealing with uncleared trades, because SIMM IM is now a direct cost embedded in derivative pricing through MVA.

SIMM calculates initial margin as a function of risk sensitivities — not portfolio MTM or notional. A portfolio with high delta (linear rate sensitivity) and high vega (volatility sensitivity) will have high SIMM IM even if the current MTM is zero. This makes SIMM fundamentally different from variation margin (which tracks MTM) and means that SIMM IM can change materially with market moves even on static portfolios. Claude helps with SIMM calculations, portfolio analysis, and MVA estimation. See XVA Trading Desk AI for the broader XVA context and XVA Explained for the MVA relationship.

SIMM Sensitivity Inputs

SIMM requires three types of sensitivities for each risk class: delta (first-order sensitivity to the risk factor), vega (sensitivity to implied volatility, for options), and curvature (second-order sensitivity, also for options). For interest rate products, delta is the DV01 or equivalent risk sensitivity to each tenor bucket on the risk-free curve and on the counterparty credit spread curve. For FX, delta is the FX spot sensitivity. For equity, delta is the equity price sensitivity.

The IR delta sensitivity for a USD IRS at tenor bucket t is the change in NPV for a 1bp parallel shift in the USD OIS curve at tenor t — the standard DV01 vector. SIMM specifies 12 tenor buckets for IR delta: 2W, 1M, 3M, 6M, 1Y, 2Y, 3Y, 5Y, 10Y, 15Y, 20Y, 30Y. For a 5Y pay-fixed USD IRS, essentially all the delta is concentrated in the 5Y bucket (with small opposing positions at shorter tenors due to float leg resetting), making the SIMM input vector very sparse in practice.

"SIMM IR delta sensitivity calculation for a USD IRS portfolio: Calculate the SIMM IR delta inputs for a portfolio of USD IRS trades: (1) pay-fixed 5Y IRS, $50M notional, current DV01 (10bps): approximately $23,500 in the 5Y bucket; (2) receive-fixed 10Y IRS, $30M notional, DV01 approximately $28,400 in the 10Y bucket; (3) pay-fixed 2Y IRS, $80M notional, DV01 approximately $14,800 in the 2Y bucket. Map each trade to its primary SIMM tenor bucket(s). Compute the net delta vector across all 12 SIMM tenor buckets. The SIMM risk weight for the 5Y USD bucket is 46bps (SIMM 2.6). Apply the risk weight and compute the within-bucket capital: K_b = sqrt(sum_k (RW_k × s_k)² + sum_{k≠l} ρ_{kl} × RW_k × s_k × RW_l × s_l). Use the prescribed within-bucket correlation matrix for IR (ρ = e^{-1% × |t_k − t_l| / min(t_k, t_l)}, capped at specified values)."
"SIMM sensitivity calculation for a cross-currency swap: We have a EUR/USD cross-currency swap: receive EUR fixed 1.8%, pay USD SOFR+40bps, €50M/USD notional, 7Y maturity. SIMM requires separate sensitivity calculations for the IR risk class (EUR and USD rate sensitivity) and the FX risk class (USD/EUR spot sensitivity). (1) Compute the EUR IR delta: the EUR fixed leg has DV01 approximately €28,000 in the 7Y EUR bucket. What is the approximate USD IR delta from the SOFR-linked floating leg? (2) For FX: the FX delta is approximately the USD NPV of the EUR/USD basis — estimate this as the initial NPV of the FX forward leg. (3) Apply SIMM risk weights: IR 7Y EUR = 46bps, FX USD/EUR = 7.4%. (4) How are the IR and FX components combined in the total SIMM IM — is there diversification benefit across risk classes?"

SIMM Aggregation and Concentration Thresholds

After computing weighted sensitivities (WS = RW × s) for each risk factor, SIMM aggregates them through a two-level hierarchy: within-bucket aggregation (combining sensitivities to different tenors within the same risk class and currency) and cross-bucket aggregation (combining across different currencies or sectors). The prescribed correlation matrices are published by ISDA and updated annually. A key feature of SIMM is the concentration risk add-on: when sensitivities in a bucket are large relative to the concentration threshold (CT), SIMM applies an additional concentration multiplier that increases IM, penalizing concentrated single-currency or single-name positions.

"SIMM concentration risk calculation for a large USD rate position: Our USD IR delta is heavily concentrated — we have $450M net DV01 in the 5Y and 10Y USD buckets combined, against a SIMM concentration threshold (CT) for USD IR of approximately $240M (SIMM 2.6). The concentration risk multiplier is: CR = max(1, sqrt(|s| / CT)). (1) Calculate the concentration risk multiplier for our 5Y and 10Y USD positions, (2) show how the concentration multiplier changes the weighted sensitivity WS_k = CR × RW_k × s_k, (3) compare the SIMM IM with and without concentration risk add-on for this portfolio, (4) what portfolio restructuring could reduce the concentration charge — netting, compression, or switching to cleared equivalents?"
"SIMM cross-currency portfolio aggregation: We have a multi-currency IRS portfolio with SIMM IR delta IM by currency: USD $3.8M, EUR $2.1M, GBP $1.4M, JPY $0.9M. These are within-bucket aggregated IM figures for each currency bucket. Cross-bucket aggregation in SIMM for IR uses prescribed cross-currency correlations: USD/EUR = 0.42, USD/GBP = 0.28, USD/JPY = 0.17, EUR/GBP = 0.50, EUR/JPY = 0.20, GBP/JPY = 0.15. (1) Compute the aggregated IR IM using the cross-bucket formula: K = sqrt(sum_b K_b² + sum_{b≠c} γ_bc × K_b × K_c), (2) how does this compare to simply summing the currency-bucket IMs, (3) what does the diversification benefit represent economically, (4) show the impact if we added a EUR/JPY cross-currency basis position that increases JPY IR delta IM by $600K."

SIMM Vega and Curvature for Options

For portfolios containing options — swaptions, caps/floors, FX options, equity options — SIMM requires additional vega and curvature inputs. Vega is the sensitivity to a 1% move in implied volatility (equivalent to the option's vega × implied vol / 100). Curvature is a second-order gamma-like term that captures the non-linearity of option payoffs under large market moves. Curvature is calculated by applying prescribed stress scenarios and measuring the second-order P&L impact. For swaption books, vega and curvature can be the dominant SIMM components — often larger than the delta IM from the underlying rate exposure.

"SIMM vega calculation for a swaption portfolio: We have a book of USD swaptions (OTC, uncleared, subject to UMR): (1) long receiver swaption, 1Y into 5Y, $100M notional, implied vol 85bps, vega $42,000 per 1% vol move; (2) short payer swaption, 2Y into 10Y, $75M notional, implied vol 90bps, vega $38,000; (3) long straddle (long payer + long receiver), 6M into 2Y, $50M each leg, net vega $12,000. SIMM vega sensitivity: VR_k = vega × sigma_k where sigma_k is the implied vol at that maturity-tenor. SIMM risk weight for IR vega: RW_vega = 21%. (1) Compute the SIMM vega weighted sensitivity for each swaption, (2) aggregate using the prescribed vega correlation matrix (same tenor structure as delta), (3) calculate total vega IM, (4) why is the curvature IM typically smaller than vega IM for ATM swaptions but can dominate for deep ITM/OTM positions?"

MVA: Pricing SIMM IM Cost Over a Trade's Life

SIMM IM is not a one-time cost — it evolves as the trade ages and as market conditions change. For a new IRS trade, the SIMM IM profile over time depends on: how the DV01 profile changes as the trade approaches maturity (it decreases), how the yield curve shape affects the net sensitivity (a steepener increases IM on belly-concentrated books), and how volatility changes. MVA prices the expected cost of funding this evolving IM over the trade's life at the bank's funding spread. For a long-dated IRS, MVA can be 10-40% of CVA in magnitude and represents the cost the dealer passes through to clients in trade pricing.

"MVA estimation for a 10Y USD IRS: We are pricing a new $100M pay-fixed 10Y USD IRS for a corporate client. The trade is uncleared (client has no clearing relationship) and the netting set triggers UMR (our bilateral IM exceeds €50M threshold). MVA calculation requires: (1) estimating the expected SIMM IM profile over the 10 years — at what rate does SIMM IM amortize as the swap approaches maturity? Assume the 10Y DV01 of $90,000 falls approximately linearly to $0 at maturity. Compute the expected SIMM IM at each annual interval using the current SIMM IR risk weight for 10Y USD of 46bps. (2) Compute MVA as: MVA = ∑_t E[IM(t)] × FundingSpread × ΔT × DF(t), assuming our funding spread is SOFR+55bps, risk-free rate 4.5%. (3) Express MVA as a running spread in bps on $100M over 10Y. (4) Compare: if this client were a clearing member and traded through CME, LCH IM would be approximately 60% of SIMM IM — how much MVA could we save by arranging clearing?"

SIMM Dispute Resolution and ISDA Governance

A key design principle of SIMM is bilateral reproducibility — both counterparties should calculate the same IM independently. In practice, small differences arise from rounding, sensitivity calculation methodology differences, and timing of recalibration updates. ISDA has a dispute resolution framework: if IM amounts differ by more than a prescribed tolerance (typically a few percent), parties must resolve the dispute within a defined timeline. ISDA also conducts annual SIMM backtesting and publishes validation results showing whether SIMM would have covered observed market moves — these are used by regulators to assess the model's adequacy.

"SIMM dispute resolution process: We have a SIMM IM dispute with a counterparty — we calculated €4.2M IM and they calculated €3.9M on the same bilateral portfolio. We have 15 business days to resolve under our CSA. Walk me through: (1) what are the most common sources of SIMM calculation discrepancies between counterparties — trade population differences, sensitivity methodology, or parameter version mismatch, (2) what data should we exchange first to identify the source of the €300K discrepancy, (3) if we cannot resolve it within 15 days, what does ISDA's dispute resolution protocol require, (4) while the dispute is pending, how much IM must each party post — the lower amount, the average, or the higher party's calculated amount, (5) what documentation do we need for our compliance records of a resolved IM dispute?"

Get free access