The fee you never see on the ticket
A perpetual future — a “perp” — is the contract most crypto traders actually use. Unlike a dated future, it never expires, so you can hold a leveraged position on BTC or ETH indefinitely. That's convenient, but it raises an obvious question: if a contract never settles against a delivery date, what keeps its price tethered to the real spot price of the coin? The answer is the funding rate, and it is the fee that quietly leaves your account every few hours whether or not the market moves your way.
Most traders watch entry, stop and target. Funding is the line they forget until they add up a month of statements and find hundreds of dollars gone with no losing trade to blame. It doesn't appear on the order ticket, it isn't a commission, and on a crowded, one-sided market it can cost more than the move you were waiting for.
Why funding exists: an anchor, not a charge
Funding is not a fee the exchange collects — it is a payment traders make to each other. Its whole job is to keep the perp's price close to the underlying spot price. When lots of people are long and the perp trades above spot, funding turns positive: longs pay shorts. That payment nudges longs to close and shorts to open, pulling the perp back down toward spot. When the crowd is short and the perp trades below spot, funding goes negative: shorts pay longs, and the pressure runs the other way.
So the sign of the funding rate is really a crowd-positioning gauge. A persistently positive rate means the market is leaning long and paying for the privilege. That's the mechanism; it is not a prediction, and a positive rate does not tell you where price goes next. It only tells you which side is currently paying to hold.
How the transfer actually works
Funding is exchanged at set intervals — commonly every eight hours, though some venues settle hourly. The rate is quoted per interval and applied to your position size, not your margin. That distinction is where leverage stings. Say the rate is 0.01% per 8 hours (a fairly ordinary number) and you hold a $50,000 long. You pay roughly $5 that interval, or about $15 a day. Modest.
Now suppose you put up $5,000 of margin at 10x to control that same $50,000. The funding is still charged on the $50,000 notional — so that $15 a day is being paid out of a $5,000 stake. Hold it a week through a hot, over-long market where funding spikes to 0.05% or 0.1% per interval, and the drag becomes real money against your capital while price has done nothing. These figures are illustrative; real rates vary by coin, venue and hour, so always read the live number on your own exchange rather than assuming a typical one.
- Positive rate, you're long: you pay every interval you hold.
- Positive rate, you're short: you get paid every interval.
- Negative rate, you're long: you get paid.
- Negative rate, you're short: you pay.
Why the overnight hold is where it bites
The single most expensive habit is holding a crowded perp through several funding windows without ever pricing what that costs. In a strong trend everyone piles onto the same side, funding climbs, and the last people in are paying the most to hold a consensus position. You can be directionally right for the day and still finish red once three funding payments and the fees are netted off — the classic case of a move that earned less than it cost to sit in.
This is also why funding is a discipline problem dressed up as a maths problem. The trader who tells themselves “I'll hold overnight and see” is often the same one who re-enters at 3am after a bad exit. If that pattern sounds familiar, we wrote about it separately in revenge trading crypto — funding just happens to put a running meter on the indecision.
Journaling the funding line
You cannot manage a cost you never record. Most traders' journals capture entry, exit and gross P&L, then quietly ignore funding — so the number that decided whether a week was green or red is missing from the review. The fix is not complicated: treat funding as a real line item, per trade and totalled per month, in dollars.
- Separate gross move from net result. If a perp trade made $300 on price but paid $180 in funding over four days, your journal should show both, not a single blurred figure. The lesson lives in the gap.
- Tag the long holds. A simple “held through funding” tag, totalled monthly, tells you fast whether overnight perp holds are actually paying you or slowly bleeding the account.
- Watch funding by market condition. Positions opened when the crowd was already heavily one-sided tend to carry the worst funding. Seeing that in your own history is more convincing than any rule of thumb.
This is exactly what PnL Book does for crypto traders: it works on your own past trades only — no tips, no signals, no calls on where price goes next. You import your Binance or Bybit history (export the closed P&L or snap the positions page), and the funding you paid stops being an invisible drain and becomes a number you can see, tag and total. Most perp accounts don't die to one bad trade; they leak, a few dollars a window, on holds nobody priced. The first step to stopping the leak is measuring it.