What do positive and negative funding actually tell you?

Funding is the most over-interpreted number in crypto. It is genuinely informative about one narrow thing and completely silent about several things people routinely read into it. This lesson draws the line precisely.

What it says directly

Mechanically, a positive funding rate means one thing and only one thing: over the last interval, the perpetual traded above its spot index, on that venue. Negative means it traded below. That is a measurement of a derivative's premium, not a survey of anybody's opinion.

The inference you may reasonably draw

A perp sustains a premium when demand for leveraged long exposure exceeds the supply of participants willing to take the other side at the index price. So persistently positive funding is evidence about positioning and crowding — leveraged longs are paying to be there, and paying continuously.

That is a real, useful, structural fact. It says something about fragility: a position that costs money to hold is a position with a clock on it, and a crowded one unwinds harder on bad news. It says nothing whatsoever about timing or direction.

Open interest is the required companion

Funding without open interest is close to meaningless, because a premium tells you about price while open interest tells you how much risk is actually on. The four-way reading is the same one the futures course applied to open interest, now with real-time data instead of a weekly-lagged report:

• Price up, open interest up, funding positive — new leveraged length entering
• Price up, open interest down — shorts closing rather than longs building
• Price down, open interest up, funding negative — new leveraged shorts entering
• Price down, open interest down — positions being closed or liquidated out

These are descriptions of what has already happened. They are read as predictions constantly, and that is the mistake.

Basis — the same information in comparable units

Dated crypto futures let you compute a cleaner version. Annualised basis = (futures price / index − 1) × (365 / days to expiry).

Worked: a quarterly future at $61,200 against an index at $60,000, with 90 days to expiry:

• Premium: 61,200 / 60,000 − 1 = 2.0%
• Annualised: 2.0% × (365 / 90) = 8.1% a year

Now that number is directly comparable to a funding rate and to a cash interest rate, which is what makes it useful: it tells you what leveraged length is costing across the whole term structure. And Unit 2 explains why it does not collapse to the risk-free rate — capturing it requires trapped capital on venues that carry real counterparty risk.

What funding cannot tell you

• Timing. Funding has stayed elevated through long advances and through sharp declines. Extremes have persisted for weeks.
• Direction. A crowded long is not automatically wrong. It is crowded.
• The whole market. Funding is venue-specific. One exchange's funding is one exchange's book, with that exchange's clientele and leverage limits.
• Who is on the other side. A large short may be a directional bet, or it may be a hedge against spot inventory, a market maker's offset, or a cash-and-carry position collecting the basis. The mechanical short and the bearish opinion are different things.

That last point is the crypto version of the lesson traditional markets teach with commercial hedgers: positioning data describes structure, not opinion.

Funding and open interest are observations. This lesson describes how to read them accurately and how not to over-read them. Nothing here is a signal, a recommendation, or advice.

Try it now

1. The table below is the far end of one venue's public funding history: Binance's BTCUSDT perpetual over the 12 months to 28 September 2026, 1,096 eight-hour readings, measured that day. The positive side never went past +0.0100%, the default interest component, which it printed 89 times; 256 readings were negative, and the most extreme of them came in separate episodes. Open interest is missing on purpose: the same venue's open-interest history refused a request reaching 200 days back, so a year of it is not public there.

2. The asset's spot price over the same window is below. Take the three most extreme funding readings in the table, locate the same dates on the chart, and Measure forward from each until the price moves meaningfully in either direction. Write down the three bar counts.

3. That gap between "crowded" and "resolved" — and the fact that its sign is not consistent across your three readings — is the entire discipline of this lesson. A crowded position is not a direction.