Dynamic Margining in On-Chain Derivatives: How Collateral and Liquidation Risk Change
Dynamic margining in on-chain derivatives means that a trading venue or protocol uses defined risk rules to assess whether pledged collateral can support an open derivatives position. Requirements may change with position size, market-wide open interest or other governed parameters. Separately, falling account equity can trigger maintenance-margin restrictions and liquidation. The crucial investor question is which rule changes first, and what happens when the market moves faster than collateral can be added or a position closed.
For institutions evaluating leveraged exposure, transparent calculations help only if the price feeds, collateral rules and liquidation mechanisms are dependable. Readers seeking the broader instrument background can consult Kenson’s overview of crypto derivatives; the analysis here concerns the margin engine itself.
A margin engine monitors the notional exposure of each position, account equity, accepted collateral and risk parameters. For a simple linear derivative, position notional equals the absolute position size multiplied by the relevant reference price. Initial margin normally determines the minimum equity needed to establish or increase exposure. Maintenance margin identifies the lower threshold below which the account or position may become liquidatable.
A compact model is: initial requirement = absolute notional × initial margin fraction; maintenance requirement = absolute notional × maintenance margin fraction. Equity generally includes eligible collateral and unrealized profit or loss, adjusted for funding, fees and venue-specific accounting. This is an explanatory model, not a universal protocol formula. Some venues use portfolio offsets, concentration add-ons, risk tiers or different account structures.
The open-source dYdX Chain margin specification gives a useful concrete model: its initial margin fraction increases as market open interest crosses defined notional bands; the documented maintenance margin fraction does not rise with that same open-interest adjustment. That matters because a new initial margin constraint can stop a position from expanding without independently causing immediate liquidation.
A risk engine also needs an acceptable pricing method. Depending on the venue, a margin check may use an oracle price, a derived mark price or additional safeguards against anomalous trades. A last traded price is not necessarily the liquidation reference. Contract size, price update cadence, oracle contingencies, collateral eligibility and authorized parameter changes all affect the calculation.
Free educational tool
How healthy is your hypothetical portfolio?
Seven questions about how a hypothetical portfolio is structured, with three sub-scores and an overall Health Score out of 100.
Educational tool. Results are not investment advice.
The word dynamic can describe several distinct changes. One protocol may increase initial margin when open interest exceeds a concentration band. Another may vary permitted leverage by asset, change collateral valuation after a depeg, or apply a new parameter following governance approval. A dashboard that says “margin updated” is therefore incomplete evidence. Institutional users should identify the triggering data, which threshold changed, who authorized the change, and whether it applies to existing exposure, fresh orders or both. Independent calculation against published rules is a stronger control than accepting a single displayed liquidation price.

Initial Margin, Maintenance Margin and Liquidation Are Different Decisions
Consider an account whose effective initial margin rises from 10% to 16% as open interest expands. If the account already holds a position, its ability to increase exposure may be restricted even while it remains above the maintenance threshold. A separate maintenance breach can authorize liquidation under the protocol’s rules. Some venues may instead apply risk-tier changes, collateral demands or close-only restrictions, so the precise sequence must be documented rather than inferred.
A margin call is a demand or notice to provide additional eligible collateral, where that process exists. Liquidation is the closing or transfer of a position under a defined trigger. Fully automated venues may liquidate without the traditional opportunity to meet a human-issued margin call. The distinction matters for both account monitoring and the timing of liquidity needs.
Leverage makes these thresholds sensitive to relatively small price movements. At five times gross leverage, a 10% adverse move in the underlying can erase roughly half the originally posted equity before funding costs and fees, assuming a simple linear position. A protocol can improve response speed but cannot promise a liquid market into which a stressed position can be sold.
Worked Example: A Leveraged Position Moves Toward Liquidation
Assume a hypothetical isolated long position of 50 ETH entered at $2,000 per ETH. The initial notional is $100,000, with $20,000 of eligible, dollar-stable collateral. The initial margin fraction begins at 10%, later rises to 16% under a documented risk-parameter change, and the maintenance fraction remains 7.5%. Assume linear profit-and-loss accounting, no funding or fees, no collateral haircut, and uninterrupted reference-price availability.
At entry, 10% initial margin requires $10,000, so the $20,000 balance provides a buffer. After the initial margin fraction rises, the same $100,000 notional would require $16,000 for opening or expanding exposure. The account still holds $20,000, so that increase alone does not establish a maintenance breach.
| Calculation | ETH at $1,800 | ETH at $1,700 |
|---|---|---|
| Position notional | $90,000 | $85,000 |
| Unrealized P&L | −$10,000 | −$15,000 |
| Account equity | $10,000 | $5,000 |
| Initial margin at 16% | $14,400 | $13,600 |
| Maintenance at 7.5% | $6,750 | $6,375 |
| Maintenance headroom | +$3,250 | −$1,375 |
At $1,800, equity is $10,000. The account is below the new $14,400 initial requirement, so additional exposure can be restricted, but it remains $3,250 above the $6,750 maintenance level. At $1,700, equity falls to $5,000 against $6,375 of maintenance requirement. The $1,375 deficit makes the position eligible for liquidation under the stated hypothetical rules.
The theoretical crossing price solves $20,000 + 50 × (P − $2,000) = 7.5% × 50 × P, giving approximately $1,729.73 per ETH. This is a simplified threshold, not a live platform liquidation quote. Funding charges, changing collateral values, partial fills, fees or different marks can move the actual trigger and final proceeds.
Notice how the risk decision changes without any new order being placed. First, the dynamic initial-margin fraction reduces the account’s capacity to expand. Next, price losses consume maintenance headroom. Finally, the position crosses a rule-based liquidation condition. The portfolio manager has to monitor all three states, because a statement that an account is “adequately margined” may conceal a ban on additional exposure or a narrow buffer against the next price move.

Cross Margin, Isolated Margin and Collateral Quality
Isolated margin confines a designated collateral allocation to one position or market. Cross margin shares eligible account equity among supported positions. Cross margin may reduce idle buffers, but one loss can consume funds supporting unrelated trades. Isolation reduces contagion between positions, although a venue’s backstop, insurance and loss-allocation mechanics still require examination.
Free educational tool
Test a hypothetical allocation against history
Choose assets and weights, set contributions and costs, and review historical outcomes, volatility and drawdowns. Historical data, not a forecast.
Educational tool. Results are not investment advice.
The Hyperliquid liquidation documentation describes liquidation tests based on account equity versus maintenance margin, order-book liquidation attempts and a backstop mechanism if ordinary execution cannot restore the required margin. It distinguishes cross and isolated positions. That is an example of a specific implementation, not evidence that all decentralized derivatives venues follow those rules.
Collateral quality can become decisive under stress. A stablecoin used for margin has issuer, reserve, market-price and redemption dependencies. If the collateral itself loses value, equity can deteriorate even when the derivatives reference asset barely moves. Non-dollar collateral may need haircuts; bridges and wrappers can add claims and operational risks. Institutions should separate the reported collateral balance from its recognized risk value and from what can actually be liquidated.
Where Automated Margining Can Fail
Price feeds can become stale or diverge during extreme volatility. A venue might pause trading, use fallback sources, reject anomalous marks or continue liquidating under imperfect information. Each choice creates different exposures. Kenson’s analysis of auditable oracle feeds provides the deeper data-governance treatment; this page focuses on how bad inputs propagate into margin decisions.
Execution risk is separate. Liquidation may require selling an increasingly large position into diminishing depth. Mark prices, posted bids and realizable exit prices can separate; transaction fees, liquidation penalties, network congestion and MEV-related ordering can compound the difference. Even when a liquidation transaction is accepted, proceeds may fail to cover liabilities. Such shortfalls may be absorbed through insurance funds, backstop liquidators, auto-deleveraging or other protocol-specific allocation arrangements.
A position might be profitable at the quoted mark and still incur material closeout costs because only a small fraction can transact at that level. Digital asset liquidity risk is particularly relevant where several accounts liquidate the same contract simultaneously. Pre-trade depth estimates should be stress-tested, not treated as standing commitments from counterparties. Where the venue uses partial liquidation, assess whether the remainder is actually returned to a sustainable collateral ratio and what happens if market depth vanishes between attempts.
Governance is another operational dependency. An upgrade, emergency action or parameter change may alter eligible collateral, margin fractions or closeout rules. A visible smart contract does not demonstrate the integrity of all administrators, dependencies or settlement outcomes. Nor does it automatically satisfy custody, reporting or regulatory requirements. The relevant legal regime depends on the product, provider, counterparty and jurisdiction. Digital asset operational risk includes the institution’s own monitoring, transaction approval and incident escalation procedures, not merely failures inside the protocol.
The 2025 BCBS-CPMI-IOSCO report on initial-margin transparency emphasizes predictability, liquidity preparedness and model responsiveness in centrally cleared markets. A May 2026 CPMI-IOSCO consultation discusses model governance and disclosures. These are comparisons for risk management, not rules automatically applicable to an on-chain protocol.
An Institutional Margin-Control and Review Framework
Before a trade, the review should identify the exact instrument and margin mode; accepted collateral and haircuts; opening and maintenance thresholds; maximum leverage; oracle and mark methodologies; funding schedule; governance powers; and whether ordinary liquidations rely on a deep enough market. A stress test should estimate how simultaneous price and collateral declines affect equity, and whether expected liquidation costs fit within available headroom.
During the position, the important figures are equity-to-maintenance coverage, additional exposure permitted under the current initial margin, collateral-concentration measures, oracle freshness, mark-to-executable-price divergence, cumulative funding and the estimated closeout cost at the position’s actual size. Alerts should distinguish warnings, close-only restrictions, margin demands and liquidation eligibility instead of labeling them all “margin calls.”
Free educational tool
What is your crypto risk profile?
Five quick questions about your mindset, time horizon and behaviour, with a reading list matched to your profile.
Educational tool. Results are not investment advice.
After a material move or liquidation, compare the reference mark with actual closeout prices, record fees and penalties, examine oracle status and transaction latency, and reconcile remaining collateral against account and settlement records. A control failure should lead to documented ownership, remediation and retesting rather than an assumption that automation worked as intended.
Digital asset risk management should also establish who is responsible for escalating a deviation between the venue’s displayed figures and the institution’s independently reproduced calculations. A temporary inability to verify collateral or reference prices can justify reducing permitted activity under internal policy. Review teams should preserve time-stamped inputs, parameter versions and exception approvals, since current on-chain state may no longer explain the rule applied at the moment of an earlier liquidation.
| Review point | Evidence or question |
|---|---|
| Risk model | Which inputs can change initial margin, maintenance margin or eligible collateral? |
| Market capacity | What executable depth exists for the potential liquidation size? |
| Data integrity | How current and independent are the oracle and mark-price inputs? |
| Loss allocation | Who absorbs deficits after ordinary liquidation fails? |
| Governance | Who may update parameters or pause contracts, and under what controls? |
| Post-trade audit | Can fills, fees, equity changes and liquidations be independently reconstructed? |
For the related but distinct challenge of managing execution impact in thin markets, see Kenson’s guide to slippage in low-liquidity trades.
What Investors Should Take Away
Dynamic margining in on-chain derivatives can make collateral rules observable and enforcement timely, but visibility does not remove market, model or operational risk. A disciplined review asks not just whether collateral exceeds a published percentage, but whether the price input is credible, margin thresholds are understood, and positions can be closed under stress without creating an unallocated loss.
Kenson Investments publishes educational analysis of derivatives infrastructure and digital asset risk management. To explore related research or ask about Kenson’s educational resources, get in touch with us.
Disclaimer: The information provided by Kenson Investments is for educational and informational purposes only and does not constitute financial, investment, legal, or tax advice, or a recommendation to buy, sell, or hold any digital asset. Digital assets involve significant risks, including market volatility, liquidity constraints, custody vulnerabilities, technological failures, counterparty exposure, and regulatory uncertainty. Their legal and regulatory treatment varies by asset, transaction, and jurisdiction. Past performance does not guarantee future results, and investors may lose some or all of their invested capital. Readers should conduct independent research and consult appropriately qualified financial, legal, or tax professionals before making investment decisions.









