Optimal Exit Time for Liquidity Providers in Automated Market Makers Journal: TBA
Philippe Bergault, Sébastien Bieber, Leandro Sánchez-Betancourt · 2025
Read Optimal Exit Time for Liquidity Providers in Automated Market Makers Journal: TBA by Philippe Bergault, Sébastien Bieber, Leandro Sánchez-Betancourt (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2509-06510.txt.
Introduction
When should someone supplying funds to a crypto trading pool pull their money out? Philippe Bergault and two co-authors study that timing question in a 2025 finance paper. In an automated market maker, providers earn trading fees but can lose value when prices move, a drawback called impermanent loss. Earlier work mostly assumed fixed deposits or exits decided from outside.
Here the exit time is part of the problem, set up as a stochastic control task where chance drives prices. The authors show the provider's value obeys a special inequality and that its solution is unique in a weak sense. Two numerical methods solve it, one based on operator splitting and one on regression. Calibrated simulations show the best exit depends on price volatility, fee levels, and how arbitrage traders and noise traders behave.
Catalog record
| CREATOR | Philippe Bergault, Sébastien Bieber, Leandro Sánchez-Betancourt |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 104 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2509-06510
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Optimal Exit Time for Liquidity Providers in Automated Market Makers Journal: TBA Philippe Bergault, Sébastien Bieber, Leandro Sánchez-Betancourt Abstract We study the problem of optimal liquidity withdrawal for a representative liquidity provider (LP) in an automated market maker (AMM). LPs earn fees from trading activity but are exposed to impermanent loss (IL) due to price fluctuations. While existing work has focused on static provision and exogenous exit strategies, we characterise the optimal exit time as the solution to a stochastic control problem with an endogenous stopping time. Mathematically, the LP’s value function is shown to satisfy a Hamilton–Jacobi–Bellman quasi-variational inequality, for which we establish uniqueness in the viscosity sense. To solve the problem numerically, we develop two complementary approaches: a Euler scheme based on operator splitting and a Longstaff–Schwartz regression method. Calibrated simulations highlight how the LP’s optimal exit strategy depends on the oracle price volatility, fee levels, and the behaviour of arbitrageurs and noise traders. Our results show that while arbitrage generates both fees and IL, the LP’s optimal decision balances these opposing effects based on the pool state variables and price misalignments. This work contributes to a deeper understanding of dynamic liquidity provision in AMMs and provides insights into the sustainability of passive LP strategies under different market regimes. Keywords: automated market makers; decentralised finance; liquidity provision; optimal stopping; Longstaff-Schwartz algorithm; variational inequalities; Hamilton-Jacobi-Bellman equation; viscosity solutions. 1 Introduction Decentralised finance (DeFi) is evolving rapidly. Since the launch of Uniswap, a flurry of innovations has emerged in this space, many of which revolve around the so-called automated market maker (AMM) technology. AMMs are decentralised trading venues where liquidity providers (LPs) and liquidity takers (LTs) interact under pre-defined trading rules; see Adams et al., 2021 ; Adams et al., 2023 ; Capponi and Jia, 2021 . AMMs such as Uniswap have become a cornerstone of DeFi, enabling users to trade assets without using traditional limit order books. Among the most prominent AMM designs, the constant function market maker (CFM) maintains an invariant of the form f(x,y)=k , where x and y denote the reserves of two assets, and f is a fixed function that determines the trading rules. A key attraction of CFMs is passive liquidity provision: liquidity providers can supply capital to the pool without the need for continuous rebalancing or active order placement, as is required in traditional limit order book markets. In return for providing liquidity, LPs earn a pro-rata share of the fees collected on trades executed in the AMM. However, LPs also face risks, most notably the so-called impermanent loss (IL), which arises due to price fluctuations between the pooled assets and external market prices. Although still in its early stages, the literature on AMMs is rapidly expanding in several key directions. A foundational contribution is provided by Angeris et al., 2020 , who analyse the core properties of CFMs; see also Angeris and Chitra, 2020 for a general multi-coin setting. In particular, they show that, in the presence of arbitrageurs, the exchange rate proposed by a CFM for small trades remains within a band around the external market price, where the width of the band is determined by the level of transaction fees. The authors also establish several fundamental properties of CFMs, such as the no-splitting and no-depletion properties, and they characterise the payoff of LPs as a function of the external asset price in the absence of fees; see also Clark, 2020 , and its extension to concentrated liquidity in Uniswap v3 in Clark, 2021 . Generalisations of AMM designs have been proposed in various directions. Cartea et al., 2024b introduce decentralised liquidity pools, generalising CFMs and offering LPs a wide range of dynamic bonding curves to improve capital efficiency, while Bergault et al., 2024a ; Bergault et al., 2024b ; Bergault et al., 2024c propose an oracle-based AMM architecture aimed at mitigating impermanent loss. More broadly, the problem of hedging impermanent loss is studied in Fukasawa et al., 2024 , where the authors analyse how LPs can reduce their exposure through dynamic trading strategies, and in Milionis et al., 2022 ; Milionis et al., 2024 where the authors introduce the now well-known concept of Loss-Versus-Rebalancing. The related notion of predictable loss is discussed in Cartea et al., 2024a , where the authors also analyse liquidity provision and trading strategies; see also Milionis et al., 2023 , and Fan et al., 2021 . Arbitrage between centralised and decentralised exchanges have been studied in Cartea et al., 2025 ; He et al., 2025 , and more generally the competition between limit order books and AMMs is discussed in Aoyagi and Ito, 2025 ; Barbon and Ranaldo, 2021 and Lehar and Parlour, 2025 . Finally, equilibrium between LPs is studied in Aoyagi, 2020 and Hasbrouck et al., 2022 . It is now well understood that the fees paid to LPs are the main mechanism to compensate for the IL that arises when arbitrageurs align the pool’s reference price and the external market price. More precisely, in the absence of fees, an agent providing liquidity in a standard CFM is exposed to a concave payoff that is strictly dominated by holding the assets outside the pool. Optimal fee structures have been studied in Evans et al., 2021 ; Cao et al., 2025 ; Baggiani et al., 2025 ; Campbell et al., 2025 , while the problem of incentive design for LPs is addressed in Aqsha et al., 2025 through a game-theoretic argument. The question of take rate – the proportion of the fees kept by the protocol – has also been studied in Fritsch et al., 2022 . While fees are ideally meant to offset IL, empirical and theoretical studies have shown that LPs often experience persistent negative returns, particularly in volatile markets; see e.g., Canidio and Fritsch, 2023 . This poses a serious challenge to the long-term sustainability of passive liquidity provision; see, e.g., Hasbrouck et al., 2025 . In this paper, we are interested in studying the optimal exit time from a CFM when the goal is to minimise IL and to maximise the fees collected. More precisely, the LP can withdraw liquidity at any time before a fixed terminal horizon T , and aims to maximise the expected value of their position, accounting for both accrued fees and the realised impermanent loss coming from the price evolution. This naturally leads to a stochastic control problem, where the exit time is treated as a control variable. Closest to our work is the recent preprint by Capponi and Zhu, 2025 , which also studies the optimal exit time from a CFM. 1 1 1 We are also aware of other concurrent efforts within this general theme. For instance, Ma et al., 2025 study allocation and exit decisions in a liquid staking protocol and an AMM. Recently, Zachary Feinstein and Marina Georgiou presented related preliminary results at the 2025 SIAM Conference on Financial Mathematics and Engineering (no preprint available at the time of writing). See also Agarwal and Gobet, 2025 . This paper provides a tractable framework with closed-form expressions that are easy to implement and to analyse. In that sense, it offers valuable structural insights into the optimal exit problem. While our work addresses