Skip to main content

​
The State of Risk Management

An exchange is only as good as its risk engine. Every major derivatives venue in traditional finance - CME, Eurex, ICE - has converged on the same conclusion: portfolio margining is the gold standard. Evaluating risk at the portfolio level, accounting for correlations and hedges, is the only way to provide capital-efficient margin without compromising safety. Until now, this has been a TradFi-only capability. Crypto exchanges either run isolated margin (each position margined independently) or a rudimentary cross-margin that still sums individual position requirements. BULK brings true portfolio margining on-chain for the first time.

​
What Makes It “True” Portfolio Margining

Portfolio margining is not just cross-margin with a different name. The defining characteristics:
  1. Correlation-adjusted notionals - A long BTC / short ETH portfolio is recognized as partially hedged. The effective notional is computed using real-time cross-asset correlations, not the sum of individual notionals.
  2. Dynamic risk surfaces - Margin rates adapt continuously to leverage, position size, market impact, and volatility regime. No static tiers. No cliff edges.
  3. Marginal risk contributions - The system computes how much each position contributes to total portfolio risk. Adding a hedge reduces your margin. Adding a correlated directional bet increases it.
  4. Hedge preservation in liquidation - When a portfolio is liquidated, the optimizer explicitly preserves hedges. It only reduces positions that actually lower portfolio risk - never unwinding the hedge leg of a spread.

​
Capital Efficiency

The difference is material. Consider a portfolio with a long BTC perpetual and a short ETH perpetual at 10x:
Margin ModelMargin Required
Isolated10% BTC notional + 10% ETH notional
Cross (sum)10% of (BTC + ETH notional)
Portfolio (BULK)10% of correlation-adjusted effective notional
With BTC-ETH correlation around 0.85, the portfolio margin requirement can be 30-70% lower than the isolated equivalent. The more hedged the portfolio, the larger the savings.

​
How CME and Eurex Do It

CME SPAN runs 16 predefined scenarios per product, scanning price and volatility shifts to find the worst-case loss. Eurex Prisma uses a filtered historical simulation with full revaluation. Both are batch processes - risk is recomputed on a schedule, not tick-by-tick. BULK computes portfolio margin continuously. Lambda surfaces update in real time with regime shifts. Correlations are live. There is no batch window, no stale risk between scans.

​
The Road Ahead

Portfolio margining is a foundation, not a ceiling. The same correlation-aware risk model that handles perpetual futures can be extended to: Multi-asset collateral - Use BTC, ETH, SOL, or other assets as margin deposits. Each collateral asset is haircut based on its volatility and correlation to the position set. A BTC deposit collateralizing a BTC-perp position gets a smaller haircut than one collateralizing a SOL-perp position. Lending and borrowing - Borrow against your portfolio’s excess margin. The risk engine already knows your true risk, so it can compute safe borrowing limits without requiring separate collateral pools. Multi-instrument portfolios - Options, spot, structured products, and perpetuals in a single margined account. The correlation model extends naturally to any instrument with a price feed. A covered call (long spot + short call) would be recognized as a defined-risk position and margined accordingly. Cross-exchange netting - Positions across regional books and multiple venues netted into a single risk profile. The goal is a single account where your entire portfolio - across instruments, collateral types, and strategies - is evaluated as one unit. Maximum capital efficiency, minimum unnecessary liquidation risk. See Margin for the full technical breakdown of how portfolio margin is computed today.