A delta-neutral position holds the asset in spot and an offsetting short in a perpetual futures contract, so price movement produces no net profit or loss. It does not follow that price movement is harmless. Only one of the two legs is margined, only that leg can be liquidated, and it takes losses whenever the price rises — while the offsetting gain on the spot leg is unrealised and frequently sitting at a different venue entirely.
The position is flat on price and emphatically not flat on margin. That asymmetry is the central operational problem in running funding capture, and the amount of collateral posted against the short leg is what determines whether a position survives a rally it was never exposed to.
Only One Leg Can Be Liquidated
The spot leg is an owned asset. It can fall in value, but nobody can close it on your behalf.
The perpetual short is a leveraged position held against margin at a venue. If equity falls below the maintenance requirement, the venue's engine closes it automatically — the mechanics are in how perpetual futures work.
The consequence is worth stating plainly, because it is the whole risk: if the short leg is liquidated, the hedge disappears and what remains is a naked long position in a volatile asset. A strategy explicitly designed to carry no directional exposure becomes a directional bet, at the moment of maximum volatility, without anyone deciding to make one.
The Asymmetry That Catches People Out
For a spot-long, perpetual-short position, a rising price is the dangerous direction.
That is counterintuitive, because a rising price is good for the asset being held. But the gain on the spot leg is unrealised and often held elsewhere, while the loss on the short leg is immediate, marked continuously, and drawn directly against posted collateral.
The portfolio is flat. The margin account is not.
A Worked Example
Consider a $100,000 notional short perpetual opened with $10,000 of collateral — ten times leverage on that leg — hedging $100,000 of spot held elsewhere.
The price rises 8%.
- Spot leg: up $8,000. Unrealised, and at another venue.
- Short leg: down $8,000. Realised against margin. Equity falls from $10,000 to roughly $2,000.
- Net position: unchanged, as designed.
- Margin position: approaching the maintenance threshold.
Another two percentage points and the hedge is liquidated, leaving a long spot position with no offset. The strategy did not lose money on the move. It lost the structure that made the move irrelevant.
How Much Buffer? The Arithmetic
How far can price move before the short leg reaches maintenance margin? Approximately:
Price move to liquidation ≈ (1 ÷ leverage) − maintenance margin rate
On a $100,000 notional short with a maintenance margin rate of 0.5% of notional:
| Leverage on short leg | Collateral posted | Approximate adverse move to liquidation |
|---|---|---|
| 10× | $10,000 | ~9.5% |
| 5× | $20,000 | ~19.5% |
| 3× | $33,300 | ~32.8% |
| 2× | $50,000 | ~49.5% |
| 1.5× | $66,700 | ~66.2% |
Illustrative arithmetic. Maintenance margin rates are tiered by venue and position size, and the figures ignore accrued funding and fees. Read the specification for the contract being traded.
Two things follow. A 10× hedge does not survive a 10% day, and crypto has plenty of those. And the relationship is not linear — most of the protection is bought moving from 10× to 3×, with diminishing returns after that.
The cost is capital efficiency: collateral posted as buffer is not deployed elsewhere, so a larger buffer means a lower return on the capital committed. That trade-off is the actual decision, and it is made before any market event, which is the only time it can be made.
Isolated or Cross, for a Hedge Specifically
Both margin modes exist on most venues, and the right answer differs for a hedge compared with a directional position.
Isolated margin ring-fences a fixed amount against the position. The maximum loss is bounded, but the position liquidates sooner because it cannot draw on anything else. For a hedge, bounded loss is less valuable than it sounds — the loss being bounded is the leg you need to keep.
Cross margin lets the position draw on the account's whole balance. The hedge survives further, which is usually what you want, at the cost of exposing the entire account balance at that venue if it goes badly enough.
For a hedge specifically, the asymmetry favours survival: the purpose of the short leg is to remain open. But cross margin only helps to the extent there is a balance to draw on at that venue, which leads directly to the next problem.
Collateral Fragmentation Across Venues
A position whose spot leg is at one venue and short leg at another has collateral in two places that cannot see each other.
The venue holding the short leg assesses margin on what is posted there. It has no knowledge of the offsetting spot position, and no interest in it. An $8,000 unrealised gain elsewhere is worth precisely nothing to the liquidation engine.
This is why collateral planning for a multi-venue strategy is a distinct exercise from position sizing. The questions are how much sits at each venue, how quickly it can move between them, and what happens if transfers slow during volatility — which they do, exactly when the transfer is needed. Venue selection and concentration are covered in what proof of reserves does and does not show.
Some venues offer unified or portfolio margin accounts in which spot holdings at the same venue can collateralise a derivatives position. Where available, that removes much of the asymmetry — the spot gain offsets the short leg's loss inside a single margin calculation. The trade is concentration: both legs, and the collateral, at one venue.
Why Topping Up During Stress Is Not a Plan
Any collateral plan that depends on adding margin during a violent move should be assumed to fail.
In the October 2025 liquidation cascade, exchange systems became overloaded and traders could not cancel orders or add collateral during the window when it mattered — described in how liquidation cascades work. Transfers between venues also slow precisely when volatility spikes.
Collateral has to be in place beforehand. Treating a top-up as a contingency converts an operational plan into a hope about exchange infrastructure during the one hour per year when it is least reliable.
What Regulated Futures Do Differently
The contrast with regulated futures markets is instructive.
At a clearing house, initial margin is a performance bond, maintenance margin is recalculated at least daily, and falling below it generates a margin call — a demand to restore the account, with a notification and restoration sequence before anything is closed. CME Group's education materials on what margin is needed set out the framework.
Crypto perpetual venues generally have no equivalent window. The engine acts when the threshold is crossed. There is no call, no grace period, and no human in the sequence.
That difference is not a criticism of either model — continuous markets arguably cannot offer a restoration window. It does mean that collateral discipline carries the entire weight in crypto that is shared with process in regulated futures, which is a reason to size buffers more conservatively rather than less.
Frequently Asked Questions
Can a delta-neutral position be liquidated?
Yes. The hedge removes directional profit and loss but the short leg is still margined at a venue. If adverse movement erodes equity below maintenance margin, that leg is closed automatically — leaving the spot leg unhedged.
Which direction is dangerous for a spot-long, perp-short position?
A rising price. The spot gain is unrealised and often at another venue, while the short leg's loss is immediate and drawn against posted collateral. The portfolio is flat; the margin account is not.
Should a hedge use isolated or cross margin?
They trade different things. Isolated bounds the loss but liquidates sooner; cross lets the position survive further by drawing on the account balance, at the cost of exposing more of it. For a hedge, keeping the leg open is usually the priority, which favours cross — but only where there is balance at that venue to draw on.
How do you calculate the liquidation price on a short perpetual?
Approximately, the adverse move to liquidation is one divided by the leverage, less the maintenance margin rate. At 10× with a 0.5% maintenance rate, roughly a 9.5% move. Venue formulas differ and maintenance rates are tiered by size, so the contract specification governs.
Why not just use low leverage on the short leg?
Lower leverage buys a larger buffer, and most of the protection comes from moving out of high leverage rather than from the last increments. The cost is capital efficiency — collateral held as buffer is not earning. That trade-off is the decision, and it has to be made before a market event rather than during one.
The Hedge Is Flat. The Margin Is Not.
The appeal of a market-neutral position is that price stops mattering. The qualification is that it stops mattering for profit and loss, and keeps mattering entirely for margin.
Which is why the operational questions — how much collateral, at which venue, in which margin mode, with what buffer — determine outcomes more than the elegance of the hedge does. The construction is the easy part. Keeping both legs open through a market that is trying to close one of them is the work. The broader strategy context is in our delta-neutral strategy overview and in the full guide to funding rate arbitrage.
This article is published for informational purposes and does not constitute investment advice or an offer to transact. All figures are illustrative arithmetic on hypothetical positions and ignore funding, fees and venue-specific margin tiers. Leveraged positions can be closed automatically with loss of posted collateral. Capital is at risk.
This article is for information only and is not investment advice. Past performance does not guarantee future results.
