Who Loses and Why? Market Microstructure of Perpetual Futures (Extended Abstract)

Abstract

This extended abstract summarizes the essence of [Ruizhe Jia et al., 2026], which asks who loses money in perpetual-futures trading, and why. We study Hyperliquid, a leading decentralized perpetual-futures venue, using the complete on-chain record of orders and trades from March 27, 2025 through June 30, 2026 for its ten most actively traded contracts. We classify wallets into algorithmic market makers, front-end traders, and algorithmic takers; front-end traders trade manually through the venue’s interface, and we interpret them as predominantly retail. For each group we then measure profit and loss trade by trade, valuing every trade against the midprice two hours later and subtracting the fees paid. On this measure front-end traders lose 1.11billion, market makers gain $208 million, and algorithmic takers approximately break even. Three channels carry the transfer and account for the whole of it. The first is voluntary trading. Front-end traders demand immediacy, placing large aggressive orders whose size scales little with displayed liquidity, and so pay the highest execution costs; favorable post-trade price movement and maker income offset part of that cost. The second is forced exits. Leverage converts adverse price moves into forced liquidations, which exit at prices far worse than any voluntary trade, and auto-deleveraging returns only part of the loss. The third is fees. Front-end traders cross the spread on most of their volume and sit in the highest-fee tiers, and so pay most of the venue’s fees. Post-trade price impact is organized by trade purpose, though group differences persist within each purpose category. These results identify immediacy, forced exits, and fees as the principal channels of wealth transfer in decentralized perpetual-futures markets.