Foreign media, citing research by MEV-X and HSE University, states that most transactions in decentralized exchanges cause the prices in the liquidity pools to deviate temporarily from those of the external markets. Such deviations continuously provide profits for arbitrage robots and MEV seekers. The article argues that if AMM continues to charge regular fees for correcting these arbitrage activities, the total value that can be allocated to the pools and arbitrageurs will actually decrease.
Research focuses on value distribution after price deviations
The study defines this portion of the profit as “price deviation from value,” which is the sum of the profits made by arbitrageurs and the fees charged by the pool from arbitrage transactions. According to this definition, the issue is not only how much fee the pool can collect, but also how much value that is ultimately released and can be captured due to the price deviation.
The article states that as long as a positive fee is charged for arbitrage correction, the arbitrage margin will be compressed. Some price corrections that could have been made will not occur; the pool does not receive this value, and external arbitrageurs also cannot obtain it, resulting in a direct reduction in the total value.
Zero fees can push the total value to its highest level.
Research simulations show that when the arbitrage fee rate is zero, the price deviation results in the highest total value. Once the fee rate increases, the total value begins to decline; when the rate rises to nearly 0.99%, arbitrageurs have little incentive to continue correcting the deviation, and the price deviation may remain within the pool, with the pool unable to generate any income from it.
The research team also tested this conclusion under various market conditions and in six different AMM structures, including Uniswap V2, Uniswap V3, Balancer, Curve, Trader Joe, and DODO. The article states that although the amount scale varies under different models, the result of "arbitrage at zero fee rate corresponding to maximum total value" remains essentially consistent.
There is an upper limit to the traditional constant product model.
For the common x·y=k constant product model of Uniswap V2, the research conclusions are more straightforward: if liquidity providers only aim to increase their own income by setting arbitrage fees, there is a clear upper limit to the value they can obtain.
The example in the text states that in scenarios of slight price deviations, when the pool sets the arbitrage fee rate to around 0.50%, the value obtained is about half of the theoretical maximum; if a more common rate of 0.30% is used, the share obtained by the pool is even lower, yet arbitrageurs can still take a considerable portion of the profit.
hooks The scheme attempts to keep arbitrage within the pool.
The article argues that the crux of the issue lies in the separation between liquidity providers and external arbitrageurs. If the pool charges no fees to external arbitrageurs, although the total value may be the highest, the profits will flow to those external parties. The alternative solution proposed in the study is for the pool to directly rebalance internally within the same transaction after a user's trade has occurred.
This approach relies on AMM and hooks. With the help of this programmable logic, the pool can immediately execute atomic internal arbitrage after retail users complete a transaction, reducing the time window for external searchers to intervene in the transaction, as well as minimizing the MEV that could be leaked to block builders through priority fee bidding.
The article also emphasizes that the exemption from handling fees only applies to the internal correction steps of the pool; the exchange handling fees for ordinary users remain unchanged. Retail traders still execute transactions at the original rates, and the regular compensation for liquidity providers also comes from user transaction fees.
