BitMEX Q2衍生品报告：资金费率α的三大来源

TL;DR

Funding rates are the periodic payment that keeps a perpetual swap (perp) anchored to the spot price of an underlying asset. When the perp trades above the index price, longs pay shorts; when it trades below, shorts pay longs.

TradFi Perps trading volume took off in Q2 2026, with funding rates on commodity perps like WTI crude oil (WTIUSDT) becoming one of the most interesting numbers on the screen. They printed funding rates near −400% annualised at the height of the US–Iran war, which forced a question worth revisiting: what actually moves the funding rate and why does the same asset pay such different funding depending on where you trade it?

This report breaks down the three drivers that answer this question on funding rate differences:

• Exchange Demographics: The audience of each exchange is different. For example, Hyperliquid works on-chain with its traders skewed to retail and degens; Binance’s is institutional and balanced. As a result, the same token would have higher funding rate on Hyperliquid, with a visible gap from Binance as institutions find it hard to onboard a DEX.
• Oracle and Index Mechanics: Commodity perps like crude oil (WTI) are priced off front-month futures, not spot markets. When that futures curve rolls in backwardation, the index marks down daily and funding is forced deeply negative — independent of sentiment.
• Margin Currency Effects: BitMEX’s two bitcoin perps differ only in collateral — bitcoin for XBTUSD and Tether (USDT) for XBTUSDT. Yet the XBTUSDT contract pays structurally higher funding because stablecoin capital leans towards long, while there is a built-in credit risk dynamic. Going long on XBTUSDT effectively creates a short position on USDT, providing valuable protection against a potential USDT depeg.

These drivers present four clear funding rate opportunities:

• BitMEX internal arbitrage: Long XBTUSD, short XBTUSDT — a market-neutral carry of roughly +4% annualised with no leverage, both legs backed by one pool of multi-asset margin.
• Long on Hyperliquid, short on Binance: Collect the structural on-chain funding premium, expressed cleanly as a single rate swap on Boros, no bridge required.
• Crude oil (WTI) backwardation arbitrage: During the rolling period, long the cheaper CME front-month future due backwardation and short the more expensive BitMEX WTIUSDT perp into the roll.
• Long crude oil funding rate on Boros: Go long the funding rate into the post-roll normalisation — receive the floating rate as it climbs off the −400% floor toward zero.

What Are Funding Rates?

A perpetual swap never expires, so it has no settlement date to pull its price back to spot. Funding rates take over that job. At each funding interval, the exchange compares the perp’s price to an index built from underlying markets. If the perp trades above the index — a positive premium — longs pay shorts; if it trades below, shorts pay longs. The payment is proportional to position size and passes directly between traders; the exchange keeps none of it.

Funding rates are constructed via a formula. In its common form, funding is the time-weighted average premium of perp price over index, plus an interest rate term for the two currencies in the pair, clamped to a cap and a floor. In short, funding measures the basis — the gap between perp and index — and anything that widens that gap raises funding. Because the premium dominates the formula, funding mostly reads order-flow imbalance: the crowded side has to bid the perp away from the index, and funding is what they pay to stay there. Figure 1 shows the composition of funding rates on the left, and how the premium term moves through a cycle on the right.

This is how funding rates work. The interesting part is that funding on one contract tells you little, but the difference between two contracts on the same token tells you a lot. Three things drive those differences: who is trading, what the index is built from, and what collateral the contract uses. The next section walks through each driver and showcases a trading opportunity for each.

What Drives Funding Rate Differences?

As mentioned previously, funding rate differences on the same token are possible, due to three primary drivers:

1. The Margin Currency — What Collateral You Post

To understand why perpetual contracts tracking the exact same asset pay different funding rates, we must examine the structural constraints placed on the traders. The clearest factor is the margin currency.

Consider the two traders who naturally show up on each book:

• The Bitcoin-margined trader: They already own Bitcoin and post it as collateral. They are often hedging or leaning short against the stack they already hold — selling some upside and protecting against a drawdown. Their presence tilts the inverse book toward shorts.
• The USDT-margined trader: They place stablecoin dry powder to work and seek leverage. USDT holds no upside with a credit risk of de-peg, so they naturally lean toward longs and more risk-seeking.

Funding is paid by the crowded side, so the long-heavy linear book (USDT-margined contracts) runs persistently hotter than the short-heavy inverse one (Bitcoin-margined contracts). Suppose the linear XBTUSDT pays +10% annualised funding and the inverse XBTUSD pays +6%. A trader who is long both contracts in equal size is paying 10% on one leg and 6% on the other; a trader who is long the inverse (XBTUSD) and short the linear (XBTUSDT) collects the 4% gap with no net Bitcoin exposure. That 4% is not a fluke of one week — over three and a half years, the inverse-minus-linear spread averages at about −3.93% and is negative in 94% of rolling three-month windows. Given the cause of the spread is the collateral currency, this is the most durable of the three drivers: there is no way to post Bitcoin and USDT as the same thing, so the two crowds never fully merge.

Historically, capturing this spread required managing two separate margin balances: Bitcoin for the inverse leg and USDT for the linear leg. This fragmentation introduced capital inefficiency and liquidation risk on a single leg during sharp market moves, despite the overall position being delta-neutral. However, BitMEX’s introduction of Multi Asset Margining removes this mechanical friction. A trader can now post a single collateral type to margin both the XBTUSD and XBTUSDT positions simultaneously. By cross-margining the two contracts, the isolated liquidation risk disappears, making it much more capital efficient to capture this behavioural spread.

2. The Venue — Who is Trading?

Holding the asset and contract design constant, funding rates still vary significantly across trading venues due to participant demographics. For example, listing a linear XBTUSDT contract on a centralised exchange (CEX) versus an on-chain decentralised exchange (DEX) yields structurally different funding environments.

Centralised venues attract institutional liquidity providers and professional arbitrageurs who continuously suppress market premiums. In contrast, on-chain venues are accessed via self-custodied wallets and are heavily populated by retail traders with a structural long bias. Consequently, funding rates on-chain remain persistently elevated compared to centralised platforms.

In an efficient market, cross-venue arbitrage should eliminate this differential. Traders would short the higher-funding venue and long the cheaper venue until the rates converge—a process that occurs rapidly between centralised exchanges due to high capital mobility.

On-chain execution, however, introduces severe operational frictions that restrict institutional capital from flattening the spread. Closing this gap requires overcoming three concrete hurdles:

• Compliance and Custody: Institutional mandates frequently prohibit trading out of hot wallets or signing transactions via self-custodied infrastructure, blocking many trading desks from participating entirely.