Perpetual Arbitrage: Why a Price Gap Is Not Locked-In Profit

Same ticker, two contracts

BTC trades at $100,000 on one perpetual exchange and $100,200 on another. Buy the first and short the second: have you locked in $200? You have opened two derivatives. You cannot normally transfer the long contract to the other exchange to extinguish the short.

A spot trade in the same transferable asset can have a delivery path for closing the gap, though execution and transfer risks remain. Two perpetuals have no shared maturity forcing their prices together. Their funding terms are part of their value. Research on perpetual pricing explicitly distinguishes funding designs and linear, inverse, and quanto payoffs.

Research: perpetual futures pricing — Ackerer, Hugonnier, and Jermann

The exit gap determines the price profit

For one BTC of matched linear exposure, buy A at $100,000 and short B at $100,200. Later A is $101,000 and B is $101,050. The long gains $1,000; the short loses $850. Combined price profit is $150: the entry gap of $200 minus the exit gap of $50.

If instead B is $102,000 while A is $101,000, the gap has widened to $1,000 and the combined price loss is $800. Add net funding and subtract trading and financing costs to obtain the full result. These examples assume both legs execute, matching contract units and a common P&L currency.

A cheap contract can be expensive to hold

Funding accrues while the spread remains open. On matched notionals and intervals, a short receiving 0.01% while the long pays 0.03% has negative net funding of 0.02%. At $10,000 per leg, that is a $2 cost for the interval, even if the entry prices looked attractive.

Different formulas, caps, sampling windows, and reference indices mean price rank alone cannot determine the funding advantage. Compare actual payment times and amounts, not unmatched headline APRs. An expected funding receipt is not a fixed receivable until the venue’s rules make it so.

Mechanics: funding formulas and settlement — Hyperliquid

Your hedge does not share a wallet

When BTC rises, the long may gain while the short’s collateral falls. The short venue cannot normally use unrealized gains at the long venue as margin. A transfer delay or withdrawal restriction can therefore liquidate one leg of a position that looked balanced in a combined spreadsheet.

Mark prices, maintenance requirements, collateral currencies, and forced-deleveraging rules can differ. A hedge must survive each account’s constraints separately. Equal dollar entries also need checking when contract multipliers or inverse payoffs make the underlying exposures unequal.

Mechanics: mark prices and liquidation — Hyperliquid

Research it as a basis trade

A study of centralized–decentralized funding arbitrage treats the position as a basis trade exposed to funding changes, imperfect hedging, execution, liquidation, and venue disruption. Its main scenarios use synthetic DEX rates, with an observed-protocol robustness check; simulated returns are not an audited live track record.

In practice, one leg can fill while the other fails; closing both can cost more during stress. A persistent spread can reflect collateral constraints or different contract economics. It is not automatically free money overlooked by everyone else.

Research: cross-venue funding arbitrage as a basis trade

Price the whole round trip

Before an experiment, specify contract units, executable entry and exit prices, both funding schedules, fees, and collateral on each venue. Stress a wider gap, a fast underlying move, one failed order, and delayed transfers. Estimate return on all reserved capital, including the buffers needed to keep the hedge alive.

The research question is whether convergence and funding can compensate for those costs and constraints. Opposite directions reduce one source of risk; they do not settle two different contracts for you.