On chain trading patterns reveal more racing than herding behavior
Herding or Racing? Decomposing Cross-Account Order-Flow Persistence in On-Chain Markets
Computational Engineering, Finance, and Science
Summary
Trades made on decentralized marketplaces often look like many different users are copying each other, a behavior called herding. The authors show that this isn’t the full story: a large part of this behavior is actually due to separate traders racing to act on price changes, not copying each other’s trades. They introduced a way to separate true herding from this racing effect and found that racing plays a big role especially in certain order books and large trades. This can help explain why people who provide liquidity to these markets often lose money.
What this means in practice
- •For market designers: Assess and reduce losses for liquidity providers by distinguishing between herding and racing in decentralized exchanges.
- •For blockchain analytics teams: Improve trading behavior models by incorporating the ratio of cross-account to same-account trade correlations to better detect arbitrage-driven price races.
Authors
Yichong Bai, Dingyue Kan
Abstract
Most of the autocorrelation in trade signs on decentralized exchanges comes from pairs of trades sent by different wallets. For Uniswap v3, this has been read as herding dominating order splitting -- the reverse of what equity markets show. We show that this herding share mixes two quantities: the probability $p$ that consecutive orders come from the same account, which is small when thousands of wallets are trading, and the ratio $κ$ of cross-account to same-account sign correlation. For any $p$, $κ$ equals one under the public-signal benchmark of Tóth et al. and zero under independent order splitting. On Uniswap v3 USDC-WETH (one week each in 2023 and 2024), the herding share is 0.65-0.86, but $κ$ is only 0.05-0.10. On a fully on-chain EVM limit order book, the herding share is similar (0.61-0.87), but $κ$ is 0.72-0.76; among Uniswap swaps above USD 100k, it is 0.34-0.52. Where $κ$ is high, cross-account co-movement is mostly a race for the price on a centralized exchange. It is concentrated in orders a few blocks apart, and a cross-fitted gradient-boosted filter that sees only Binance data removes about three quarters or more of it. This race flow accounts for most of the losses of liquidity providers. We recommend reporting $κ$ alongside the herding share and modeling on-chain order flow as splitting traders plus arbitrageurs who respond to the centralized-exchange price.