Direct answer
When a 10,000 USDT long expects to pay 2 USDT at the next settlement (+0.02%), closing beforehand as a taker costs 5 USDT in fees alone and reopening afterwards another 5 USDT; even trimming to 4,000 USDT costs 3.80 USDT. On fees alone, skipping a settlement pays only when the rate exceeds 0.10% per interval, about ten times the common +0.01%.
Four comparable choices
Choice one holds through settlement; two exits fully beforehand; three exits and reopens afterward; four reduces only part of notional. Compare all under the same expected price path.
Comparing next funding with zero omits exit and re-entry costs. If a 10,000 USDT long expects to pay 2 USDT while one taker exit costs 5 USDT, leaving solely to avoid funding may not be economical.
Hidden friction of reopening
If the strategy still needs the same exposure after settlement, reopening adds another fee and may execute at a worse price. A 10 USDT round trip plus slippage often exceeds a 2 USDT payment.
If the thesis is already invalid or the planned exit was before settlement, funding is secondary. Do not extend a failed trade for the sake of this comparison.
The four choices, priced side by side
Take a 10,000 USDT long, a predicted next rate of +0.02% and a 0.05% taker fee. Holding costs 2 USDT of funding; closing to avoid those 2 USDT costs 5 USDT for the exit alone.
Reducing sits in between: cut to 4,000 USDT and the next payment falls to 0.80 USDT, while the fee on the 6,000 USDT reduced is 3 USDT. The smallest number belongs to doing nothing.
| Choice | Next funding | Fee now | Fee to reopen | Quantifiable total |
|---|---|---|---|---|
| Hold the position | 2.00 USDT | — | — | 2.00 USDT |
| Close before settlement, stay out | — | 5.00 USDT | — | 5.00 USDT |
| Close before, reopen after | — | 5.00 USDT | 5.00 USDT | 10.00 USDT |
| Reduce to 4,000 USDT | 0.80 USDT | 3.00 USDT | — | 3.80 USDT |
How high the rate must be to skip one interval
On fees alone, skipping an interval pays when position × rate > exit fee + re-entry fee. Divide both sides by the position and the threshold no longer depends on size.
All-taker, the threshold is 0.10% per interval — roughly ten times a common +0.01%. Posting both the exit and the re-entry brings it down to 0.04%, but only if both actually fill as maker, which in the final minutes before settlement is rarely something to count on.
| Exit / re-entry execution | Threshold per interval |
|---|---|
| Taker / taker | 0.10% |
| Maker / taker | 0.07% |
| Maker / maker | 0.04% |
Reducing size is continuous
The choice is not only all-in or all-out. Reducing notional from 10,000 to 4,000 USDT also reduces the next funding amount to roughly 40% while retaining some exposure.
The reduction still has fee and slippage, and the remaining stop, liquidation distance and margin require recalculation.
Decision table
List next funding, immediate fee, expected slippage, re-entry fee, price risk and strategy consistency for each choice. Quantifying the first five is not enough if the strategy itself is invalid.
Use the contract countdown, estimated rate and live account notional, then reconcile from funding history after settlement. An estimated rate is not a final settled rate.
Official sources and calculation boundary
That funding is computed only at settlement, and that a position closed beforehand neither pays nor receives for that interval, follow OKX's funding FAQ; fee calculation comes from its futures fee documentation. The +0.02% and 0.05% here are example values fixed so the four choices can be compared side by side.
Open the matching calculatorNext checks in this series
Frequently asked questions
Does closing one second before settlement guarantee no funding?
Do not rely on boundary timing. Execution, system clocks and settlement processing create risk; follow exchange rules with adequate buffer.
Must a position close whenever funding is high?
No. Compare execution friction, price risk and strategy validity; funding is one input.
Should I add the size back after settlement?
If you want the original exposure back after settlement, count that fee too: reducing by 6,000 and adding 6,000 back costs 6 USDT in two trades, already more than the 2 USDT of funding you avoided.
Does the question reverse when funding is negative?
Yes. With a negative rate a long is paid at settlement, so closing beforehand forfeits that income; the question becomes whether holding one more interval is worth it for the payment, again weighed against fees and price risk.