Perpetuals and the funding rate
A payment appeared against your position overnight. You didn’t trade, nothing closed, and the amount doesn’t correspond to any fee on the schedule.
That’s funding, and it’s the mechanism that makes a perpetual contract work at all. It goes to or from other traders rather than to the venue, and it’s the price of a design decision made when the instrument was invented.
The problem funding solves
A conventional futures contract has an expiry. At expiry it settles against the underlying, and that settlement is what anchors it: however far the future drifts from spot beforehand, convergence is guaranteed on a known date, and arbitrage prices that in continuously.
A perpetual has no expiry. It never settles, so there is no convergence event and nothing structural stopping it from drifting arbitrarily far from the spot price of the thing it references.
Funding is the substitute. Instead of one convergence at a known date, there is a continuous economic pressure applied at intervals, pushing the contract back toward the index.
How it works
At each funding interval, holders of positions on one side pay holders on the other, in proportion to position size.
When the perpetual trades above the index, funding is typically positive, and longs pay shorts. Holding a long costs something and holding a short earns something, which makes shorting more attractive and longing less so, and pushes the contract price down toward the index.
When it trades below, funding is typically negative, and shorts pay longs. Same mechanism in reverse.
Three properties that are commonly misunderstood.
The venue is not a party. Funding is a transfer between position holders. The venue typically facilitates it and does not collect it — though whether any spread or component of the calculation accrues to the venue is one of the things that varies.
It’s charged on position notional, not on margin. A leveraged position pays funding on the full size, which is why funding can be a meaningful cost against posted capital on a highly leveraged position.
It applies to open positions at the funding moment, not to trading activity. A position opened and closed between intervals pays nothing. A position held across one pays regardless of whether it was profitable, and regardless of when in the interval it was opened — snapshot timing rather than time-weighting is the common convention, though not a universal one.
The rate is typically derived from two components: the observed premium of the contract over the index, and an interest-rate component reflecting the difference between the two currencies in the pair. The exact formula, the interval length, any caps on the rate, and whether the snapshot is instantaneous or averaged are all venue decisions and all differ.
Three prices
Perpetual venues quote several prices that are not the same number, and conflating them is the source of the most consequential surprise in the instrument.
Last traded price. The price of the most recent trade on this venue’s book. A single fact about the past, and manipulable in a thin book by a single aggressive order.
Index price. A composite of spot prices across multiple external venues, usually with rules for excluding outliers and handling a source going offline. It represents the underlying, not this contract.
Mark price. The venue’s fair-value estimate for the contract, typically built from the index plus a damped measure of the contract’s own basis. Its purpose is to be resistant to manipulation and to momentary wicks in the local book.
Unrealised profit and loss, margin ratios, and liquidations are generally referenced to the mark price, not the last traded price. The reason is protective: if liquidations keyed off the last trade, a participant could trigger cascades by pushing a thin book briefly through a level. Marking to a manipulation-resistant composite removes most of that.
The practical consequence is the one that surprises people in both directions. A visible wick to a price on the venue’s own chart may not liquidate a position, because the mark price never went there. And a position can be liquidated at a moment when the local last price looks safe, because the mark price — driven by the external index — went further than the local book did.
The mechanism
THE MECHANISM — funding and marking
· Perpetual trading above the index
→ funding typically positive.
Longs pay shorts.
· Perpetual trading below the index
→ funding typically negative.
Shorts pay longs.
· Position open at the funding snapshot
→ pays or receives on full
notional, regardless of P&L
· Position opened and closed between
snapshots
→ no funding. Snapshot timing,
not time-weighting, is the
common convention.
· Funding recipient
→ the other side of the market,
NOT the venue.
· Liquidation reference
→ generally mark price, not last
traded price. A local wick may
not liquidate; an index move
may.
· Interval length, formula, caps, index
composition, mark construction
→ VENUE-SPECIFIC. Every element
of this varies.
Worked example
Illustrative figures, synthetic throughout.
Suppose a position of 1 unit at a mark of 40,000 — notional 40,000 — and a funding rate for the interval of +0.01%.
A long pays 40,000 × 0.0001 = 4, transferred to shorts. A short receives 4.
If the position was opened with 4,000 of margin at 10× leverage, that 4 is 0.1% of posted capital for one interval. Across many intervals in a persistently positive-funding regime, that compounds into a cost that is not small relative to the margin, while remaining tiny relative to the notional.
Which is the whole point about leverage and funding: the rate looks negligible against the position and is not negligible against your capital.
What funding is not
Not a fee. Fees go to the venue and are charged on trading activity. Maker-taker is a separate bill entirely, and both apply.
Not interest on borrowing. It resembles a carry cost and it isn’t one — the direction depends on where the contract trades relative to the index, so it can pay you.
Not a signal. Funding tells you the contract has been trading above or below the index. This site does not attach any predictive interpretation to that, and treatments that do are outside what microstructure supports.
The failure mode
Funding is a cost or a credit that accrues without any action from you, and it is easy to leave out of an assessment of what a position costs to hold.
The specific way it bites: a position held through a persistent funding regime accumulates payments that are invisible in the entry and exit prices. Two traders with identical entries and exits can have materially different outcomes purely from how long they held and what funding did during it.
And the marking mechanism has its own failure mode in the other direction. Because liquidation references the mark price rather than the local book, the price you’re watching is not the price that determines your liquidation. A position can be closed out on the basis of a number that never appeared on the venue’s own trade tape — behaving exactly as designed, and entirely unlike what the chart in front of you suggested.