Funding Rate Arbitrage

A perpetual future never expires, so nothing forces it back to the spot price. Exchanges bolt on a mechanism instead: every eight hours, whichever side is trading at a premium pays the other. When the perp trades above spot, longs pay shorts. That payment is the funding rate, and it is the entire reason this trade exists. Buy the coin, short the perp against it, and you own no price risk at all — you own a stream of funding payments made by leveraged longs who want exposure more than you do. It is the crypto version of cash-and-carry, and for long stretches it has been the most reliable yield in the asset class.

Where the yield comes from

Funding is quoted per eight-hour period, which makes it easy to misjudge. Three payments a day, 365 days a year, multiplies small numbers into large ones.

Funding per 8h	Per day	Simple annualised	Typical regime
+0.01%	0.03%	10.95%	the floor most venues default to
+0.03%	0.09%	32.9%	steady uptrend, retail levered long
+0.05%	0.15%	54.8%	crowded, weeks at a time
+0.10%	0.30%	109.5%	blow-off, days at most
−0.02%	−0.06%	−21.9%	bear market — you are the payer

Worked example: the base trade

Take $1,000,000 and BTC at $100,000.

• Buy 10 BTC spot on Venue A: $1,000,000.
• Short 10 BTC of the perp on Venue B, posting $250,000 of margin.
• Net delta: zero. Whatever BTC does, one leg gains what the other loses.

Hold for 90 days at an average +0.01% per period — the floor, deliberately conservative:

• Funding collected: 0.03% a day for 90 days is 2.70% of notional = $27,000.
• Trading costs: taker fees of 5bp a side, in and out, on two legs — four fills at $500 each = $2,000.
• Net: $25,000.

But the capital employed is $1,000,000 of spot plus $250,000 of perp margin. $25,000 on $1,250,000 is 2.00% over 90 days, or 8.1% annualised — not the 10.95% headline. The margin you must idle against the short is a permanent tax on the trade, and it is the first thing beginners forget.

Worked example: how it actually goes wrong

Keep the same book: $250,000 of margin against a $1,000,000 short perp on Venue B, spot sitting on Venue A.

BTC rallies. At $120,000 the short is down $200,000. At $125,000 the short is down $250,000 and the margin is gone. The spot leg is up an identical $250,000 — but it is on a different exchange, and moving collateral across venues takes minutes on a good day and hours during a squeeze when everyone else is trying to do the same thing.

If Venue B liquidates you at $125,000 you crystallise a $250,000 loss and are left holding 10 BTC completely unhedged, at the top of a squeeze, in a trade you entered because it had no price risk. Bitcoin has printed 25% weekly moves repeatedly — down in March 2020, up in February 2021 — so this is not a tail you can wave away.

Carry a 50% buffer instead and you survive a 50% rally, but capital rises to $1,500,000 and the same $25,000 becomes 6.8% annualised. That trade-off, not the funding forecast, is the actual decision.

What erodes the edge

Funding flips. The stream is not a coupon. Through most of 2022 funding sat negative on the major venues, turning the carry trade into a carry cost for anyone who did not unwind.

Venue risk swamps market risk. In November 2022 FTX froze withdrawals and filed for bankruptcy. Every delta-neutral book with a leg there learned that a perfect hedge on a failed exchange is worth zero. Counterparty exposure is the position; the BTC is incidental.

Auto-deleveraging. When a cascade exhausts an exchange's insurance fund, the venue force-closes profitable positions to balance the book. Your winning short is exactly what gets taken. The October 2025 liquidation cascade, the largest crypto has recorded, hit basis books through mechanics like this rather than through price.

Institutionalisation. Since the US spot bitcoin ETFs launched in January 2024, the same carry has been available through regulated CME futures to balance sheets far larger than a crypto fund's. More capital chasing it means funding spends far more of its time pinned at the 0.01% floor and far less at 0.05%.

In interviews

State the mechanism in one line — perps have no expiry, so funding is the tether — then annualise a per-period rate out loud, because that arithmetic is half the question. Walk the P&L: funding on notional, fees on both legs, return on capital including idle margin. Then pivot unprompted to the risk, and make the specific point that the danger is not price but the asymmetry between venues: the losing leg is margined and the winning leg is not liquid enough to rescue it in time. Finish by naming counterparty and auto-deleveraging risk. A candidate who calls this a riskless yield has failed the question.