This article is for Options trading through our demo environment via API.
This document provides a comprehensive technical overview of the Options Portfolio Margining System, describing its methodology, margin calculations and risk management strategies.
Portfolio margining is a risk-based margin methodology that determines margin requirements based on the overall risk of a portfolio, rather than on predefined strategies. This approach aims to align margin requirements more closely with the actual risk exposure of the combined positions within the portfolio.
Therefore, the portfolio margining system offsets risks across different instruments with similar underlying assets. It improves capital efficiency compared to traditional margining, which sums up individual margin requirements. However, it introduces more complexity in margin calculation.
The margin requirement is calculated based on five primary risk factors:
Non-delta Market Risk – Measures the worst-case scenario P&L under price and volatility stress tests following the SPAN methodology.
Absolute Options Delta – Accounts for liquidation and market impact risk.
Net Portfolio Delta – Captures potential delta-hedging costs during liquidation.
Futures Positions Margin – Ensures consistency with multi-collateral futures margining.
Cross-Asset Netting – Allows margin reductions based on asset correlations.
This component calculates the worst-case portfolio P&L under multiple price and volatility shock scenarios. The methodology follows the SPAN (Standard Portfolio Analysis of Risk) model but with enhancements to handle options (an important difference is that we consider only the delta-hedged portfolio):
23 different market scenarios simulate price movements as an example between -15% to +15%, with additional extreme shocks at -45% and +45%.
Volatility adjustments based on an option’s historical IV behavior, and forward-looking risk estimates.
Theta Decay Impact: Additional risk multipliers applied for options with negative theta i.e. those that lose value over time, by shocking the time to expiration and bringing it forward in time.
Delta-Hedged Stress Testing: By assuming a delta-hedge at the current prices for the price shocks, only the residual risk (high order greeks) is considered.
Absolute delta measures the total directional exposure of a portfolio. This ensures large positions are margined adequately to reflect liquidation/market impact risk. Although a portfolio can have minimal market risk according to the Non-Delta Market Risk, a high delta portfolio will have a liquidation risk. Therefore, the delta of the options’ position is calculated:
Where MaintenanceMarginFactor is given here. A factor of 2 is used as a buffer.
Option | Delta | Position | Underlying price |
|---|---|---|---|
A | 0.5 | 100 | $50 |
0.5 | -0.3 | 150 | $40 |
Compute absolute notional delta exposure:
(0.5 x 100 x 50) + (-0.3 x 150 x 40)=4300
Suppose the maintenance margin factor (mm_factor) is 0.01:
AbsOptionsDelta = 4300 x0.01 x2 = 86
Net delta represents the overall directional bias of the portfolio, considering both options and futures positions. This metric is crucial for determining hedging costs before liquidation.
The methodology:
Considers both options and futures in the portfolio.
Calculates net delta exposure, incorporating futures hedging.
Uses the smallest unhedged exposure to determine risk.
Converts delta exposure to notional value using the index price.
Applies a dynamic maintenance margin factor to compute the required margin.
Instrument | Type | Delta | Position | Index Price |
|---|---|---|---|---|
A | Option | 100 | 100 | $50 |
B | Option | 200 | 200 | $50 |
C | Future | -80 | -80 | $50 |
Calculate Options Delta:
options_delta = (0.5 x 100) + (-0.3 x 200) = 50 - 60 = -10
Calculate Futures Delta:
underlying_delta= -80
Determine Net Delta Exposure:
min_net_portfolio_delta=
=minabs(options_delta), abs(options_delta + underlying_delta )=
=min10,abs(-10+(-80))=10
min_net_portfolio_delta is bounded by abs(options_delta) since we only want to consider futures deltas that are offsetting options deltas.
Convert to Notional Exposure:
min_net_portfolio_delta_notional = 10 x 50 = 500
Apply Maintenance Margin Factor (ex. mm_factor = 0.01):
Net Portfolio Delta = 500 x0.01 = 5
Futures margining follows the multi-collateral margining system to ensure that margin requirements for futures remain consistent for the clients that trade only futures; we use the existing margin system for multi-collateral futures.
Margin requirements for futures are summed with the margin requirements calculated for options.
Note: The multi-collateral margining system is slightly different from the one for single collateral contracts. For single collateral, margin requirements are based on the mark price instead of the average entry price. This approach for multi-collateral contracts was chosen since it was deemed easier for clients to understand, since initial margin is fixed for a given size and average entry price.
The worst of the summed losses per scenario across all options.
The sum of the worst loss per asset across all scenarios (the stricter).
The system interpolates between these methods to balance risk accuracy and capital efficiency. And this interpolated value will be considered the portfolio non-delta market risk. The parameter itself will be set by the administrator and will be between 0 and 1.
Note: This number expresses only the correlation between BTC and ETH, since these are the underlyings of the options that are listed and it is expected to have for some time. If further options are added the methodology needs to be revisited to handle 3+ distinct underlyings.
Scenario | BTC Loss ($) | ETH Loss ($) |
|---|---|---|
1 | -1,000 | -2,000 |
2 | -500 | -2,500 |
3 | -1,500 | -1,500 |
4 | -2,500 | -500 |
The losses for each scenario are the result of the sum of the loss of each asset. It assumes that all losses in each scenario are additive, meaning BTC and ETH losses are fully realized together.
Scenario | Total loss |
|---|---|
1 | -3,000 |
2 | -3,000 |
3 | -3,000 |
4 | -3,000 |
The worst-case total across all scenarios is -3000.
Here, we look at the worst loss for each individual instrument across all scenarios:
Worst BTC loss across all scenarios: -2500 (Scenario 4)
Worst ETH loss across all scenarios: -2500 (Scenario 2)
The total margin requirement is -2500 + (-2500) = -5000
However, this does not assume BTC and ETH hit their worst losses in the same scenario.
After calculating the above risk factors for individual positions, the final margin requirement is determined at the portfolio level by integrating these components in a structured manner:
OptionsMaintenanceMargin = max(CrossAssetNettedMarketRisk, AbsOptionsDelta)+ NetPortfolioDelta
We take the max of the two above so that we are still bound by liquidation risk in the case of a carefully hedged portfolio.
OptionsInitialMargin = OptionsMaintenance x MarginOptionsIMarginFactor
PortfolioMaintenanceMargin = OptionsMaintenanceMargin + FuturesMaintenanceMargin
PortfolioInitialMargin = OptionsInitialMargin + FuturesInitialMargin
The OptionsIMarginFactor is defined by the administrator.
In the case of long-only option portfolios, the options initial and maintenance margin cannot be more than the mark price, since this is the maximum loss.
The engine calculates an initial margin taking into account the open orders and assumes they are executed in order to prevent the clients from sending orders which, if executed, will increase the margin required more than the initial margin.
The mechanism ensures that orders which could increase risk are adequately margined before execution. For portfolios with active open orders, the system simulates margin impact assuming all open orders are filled and add any negative PnL and fees. Therefore, per option, the required margin for orders is:
marginForBids = marginImpactAllBidsFilled+min(pnlBidsFilled,0) + bidsFilledFee
marginForAsks = marginImpactAllAsksFilled + min(pnlAsksFilled,0) + asksFilledFee
marginRequiredForOrders = max(marginForBids, MarginForAsks)
We take the safe side and therefore the highest margin needed. Bid, ask orders cannot be fulfilled at the same time.
The open order margin calculation is described below:
1. Compute Initial Margin for Orders per Option
For each option opt, the initial margin impact is computed for both bids and asks, considering:
Impact if filled
Unrealized loss (negative PnL only)
Fees associated with the orders
Therefore, the initial margin of the open orders is:
IM[orders,opt] = max(max[of bid&ask orders](IM[impact](orders) + min(PNL[impact](orders),0) + Fees[impact](orders)),0)
Note: The orders at bids and asks cannot be filled at the same time. This is the reason the max of each is taken.
2. Sum margin for all option orders unless the client is a Market Maker where only the N largest margin-impacting orders are summed.
3. The total IM for all orders is the sum of those calculated above (options) and the one for futures.
When the user is placing an order, the total margin for open orders is calculated assuming the order is placed. If the difference between this value and the value before placing the order (IM[orders, total]) is less or equal than the usable margin, then the order may be placed.
Check if the total initial margin for open orders increases if the order is executed. If it is not then place the order.
The margin impact is calculated:
max(mm_impact_all_bids_filled, mm_impact_all_asks_filled)
If the calculation above produces a negative margin impact, then this means that the order reduces risk and the system allows margin buffer consumption:
usableMargin = margin_equity - current_mm where margin_equity is the collateral value (including haircuts) + unrealized PnL.
If the calculation above produces a positive margin impact, then this means that the order increases risk and the system uses: usableMargin = availableMargin
If a user's available margin falls below zero and they have risk increasing orders, then the system will start to cancel orders in an attempt to bring their available margin above zero.
If the instrument is a future and has non-closing orders (i.e., orders that do not close positions/increase exposure), then all orders for that instrument are canceled.
Unlike Futures, due to the non-linearity of options, an order executing may result in reduction of the portfolio’s total risk by closing positions (ClosingOrders). Therefore, if a group of buy/sell orders decreases risk, the buy/sell orders are cancelled otherwise not.
Iterate Over All Instruments: Each instrument can be a future or an option.
Handling Futures: Cancel All Non-Closing Orders
If the instrument is a future and has non-closing orders (i.e., orders that increase exposure), then all orders for that instrument are canceled.
Handling Options: Selective Order Cancellation
Group all buy and sell orders together in two groups (sell and buy) and then calculate whether each group reduces the margin or not. If the margin is not reduced, cancel all orders from that group (even those that decrease the margin).
Note: Please keep in mind that open orders do not increase margin while still open. The reason we cancel all those that can potentially increase the required margin if executed, is that we do not want the client to have positions that require margin.