💻 computer science

Characterizing Path-Independent Fees: A Route to Zero Impermanent Loss in CPMMs

This paper characterizes the specific fee structures required to achieve path independence in Constant Product Market Makers, derives the resulting pool dynamics, and demonstrates that while zero Impermanent Loss is attainable for specific initial states through tailored fee functions, no universal fee can eliminate it across all scenarios.

Original authors: Andrey Voronin, Roman Vlasov, Vladimir Gorgadze, Andrey Seoev, Yury Yanovich

Published 2026-05-01

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Andrey Voronin, Roman Vlasov, Vladimir Gorgadze, Andrey Seoev, Yury Yanovich

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine a digital marketplace where people trade tokens without a central boss. This marketplace is run by a smart contract called a Constant Product Market Maker (CPMM). Think of this contract as a giant, self-balancing seesaw. On one side sits Token A, and on the other sits Token B. The rule of the seesaw is simple: the product of the two sides must always stay the same number. If you add more Token A, the contract automatically removes some Token B to keep the balance.

People who put their tokens into this seesaw are called Liquidity Providers (LPs). They hope to earn money from trading fees. However, they face a tricky risk called Impermanent Loss. Imagine the price of Token A suddenly skyrockets. The seesaw automatically shifts to sell your Token A and buy more Token B to keep the balance. When you eventually take your money out, you have fewer of the expensive Token A and more of the cheap Token B than if you had just held them in your pocket. That difference is your "loss."

This paper tackles two big problems with how these marketplaces currently handle fees and risks.

1. The "Splitting" Problem (Path Dependence)

Currently, when you trade, the contract charges a small fee (like a toll). Usually, this fee is a fixed percentage. The paper points out a hidden flaw: it matters how you split your trade.

The Analogy: Imagine you are walking across a bridge that charges a toll. If you walk across in one big step, you pay one toll. But if you take 100 tiny steps, the current system might charge you a toll for every single step. Even if the total distance is the same, the final cost (and your final position) changes depending on how you broke up the journey.
The Paper's Fix: The authors figured out the exact mathematical rule for a fee system where it doesn't matter how you split the trade. Whether you trade in one giant chunk or a million tiny crumbs, the result is exactly the same. They call this Path Independence.
The Secret Rule: To make this work, the fee cannot just be a random number. It must be tied strictly to the "size" of the seesaw (the total value of the pool). If the pool grows, the fee rate can change, but it must change in a very specific, predictable way based only on that total size, not on the specific mix of tokens.

2. The "Zero Loss" Dream

The second question the paper asks is: Can we design a fee system that completely eliminates Impermanent Loss for the people providing the money?

The Good News: Yes, but with a catch. The authors created a special "magic fee" formula. If the pool starts at a specific size, this fee adjusts dynamically.
- The Analogy: Think of this fee as a smart insurance policy. If the market moves slightly, the fee is tiny (almost free). But as the market moves further away from the starting point, the fee gets higher and higher. This extra money collected from traders is used to perfectly pay back the Liquidity Providers for their "losses."
- The Result: For a pool starting at a specific point, this system can mathematically guarantee that the Liquidity Provider ends up with exactly the same value as if they had just held their tokens in their pocket, no matter how much the price swings.
The Bad News (The "No Free Lunch" Theorem): The paper proves you cannot have a universal magic fee that works for every starting point at the same time.
- The Analogy: You can design a pair of shoes that fits one person's foot perfectly. But you cannot design one single pair of shoes that fits every human foot perfectly at the same time.
- The Conclusion: If you want zero loss, the fee system must be "state-aware." It needs to know exactly where the pool started. If the pool drifts too far from that starting point, the "zero loss" guarantee breaks, and you have to reset the system or accept some loss.

Why Does This Matter?

The paper provides a blueprint for building better digital marketplaces:

Predictability: By using "Path Independent" fees, developers can ensure that the outcome of a trade is the same no matter how complex the transaction is. This makes the system safer and easier to build other apps on top of (a concept called "composability").
Fairness: The "Zero Impermanent Loss" design offers a way to protect investors, but it requires a dynamic fee that gets smarter as the market moves.
Reality Check: It proves that while we can protect investors in specific scenarios, we cannot create a single, perfect fee system that works for every possible market condition simultaneously.

In short, the authors have mapped out the mathematical rules for a fairer, more predictable trading system, showing us exactly how to charge fees so that the math works out perfectly for everyone involved—provided we accept that the rules must adapt to the specific situation.

1. Problem Statement

The paper addresses two critical issues in Constant Product Market Makers (CPMMs) like Uniswap V2:

Path Dependence: In standard CPMMs with fee reinvestment, the final state of the pool (reserves and invariant $k=xy$) after a trade depends on how the trade is fragmented (split into smaller transactions). This complicates risk assessment for Liquidity Providers (LPs), creates arbitrage opportunities for strategic fragmentation, and undermines composability guarantees in DeFi.
Impermanent Loss (IL): LPs face losses when token prices diverge from the initial deposit ratio. While fees compensate for this, standard fixed-fee models do not perfectly offset IL, and the relationship between fee structure and IL is not fully characterized.

The authors ask: What functional form of fees guarantees path independence, and within that class, can we design fees that eliminate Impermanent Loss?

2. Methodology

The authors employ a rigorous mathematical approach combining:

Differential Geometry & Calculus: Modeling pool dynamics as a system of Ordinary Differential Equations (ODEs) where the trade size is the independent variable.
Integrability Conditions: Analyzing the exactness of differential forms to derive necessary and sufficient conditions for path independence.
Parametric Construction: Deriving specific fee functions that satisfy the zero-IL condition for a given initial state.
Numerical Simulation: Validating theoretical derivations through controlled experiments comparing path-independent designs against standard Uniswap V2 implementations.

3. Key Contributions & Results

A. Characterization of Path-Independent Fees (Theorem 1)

The paper proves that a fee structure is path-independent if and only if the combined fee factor $\alpha = 1 - \gamma_1\gamma_2$ (where $\gamma_1, \gamma_2$ are input/output fee fractions) depends solely on the current pool invariant $k = xy$.

Mathematical Form: $\alpha(x, y) = \Phi(xy) = \Phi(k)$ .
Implication: The fee rate must be constant along the hyperbolas of the invariant curve. If the fee depends on the price ratio ( $y/x$ ) or individual reserves, the outcome becomes path-dependent.
Flexibility: Protocol designers can still vary the split between input and output fees ( $\gamma_1$ vs $\gamma_2$ ) as long as their product satisfies the condition.

B. Closed-Form Exchange Formula (Theorem 2)

For any path-independent fee function $\Phi(k)$ , the authors derive a system of ODEs governing the pool dynamics. They solve this to obtain a closed-form integral formula:
$G(k_f) - G(k_0) = \ln\left(1 + \frac{\Delta x}{x_0}\right)$
Where $G(k) = \int \frac{dk}{\Phi(k)k}$ .

Significance: This allows protocol designers to calculate the final pool state and invariant growth without simulating every micro-step of a trade, provided the fee function is known.

C. Zero Impermanent Loss Construction (Theorem 3)

The authors construct a parametric family of fee functions, $\Phi_{k_0}(k)$ , designed to achieve zero absolute Impermanent Loss for a specific initial invariant state $k_0$ .

Mechanism: The fee is zero at the reference state ( $k=k_0$ ) and increases monotonically as the invariant $k$ deviates from $k_0$ .
Result: The accumulated fee revenue exactly offsets the loss from rebalancing for any trade size $\Delta x$ starting from $k_0$ .
Behavior: For small deviations, $\Phi_{k_0}(k) \approx 2\sqrt{k/k_0 - 1}$ .

D. Impossibility of Universal Zero-IL Fees (Theorem 4)

The paper proves a negative result: No single fee function $\Phi(k)$ can eliminate Impermanent Loss for all possible initial states $k_0$ simultaneously.

Reasoning: The fee required to zero-out IL for a specific $k_0$ is a strictly decreasing function of $k_0$ . Therefore, a fee function that works for one initial state will fail for another.
Conclusion: Zero-IL mechanisms must be state-aware (recomputed when liquidity is added/removed) or target a specific reference range.

4. Experimental Validation

The authors validated their theory through numerical experiments:

Path Independence: Simulations showed that path-independent designs ( $\Phi(k)$ ) achieve machine precision ( $\epsilon \approx 10^{-16}$ ) regardless of trade fragmentation ( $N$ ). In contrast, standard Uniswap V2 (discrete fee compounding) showed small errors ( $\sim 10^{-5}$ ) that scaled with fragmentation, confirming the theoretical necessity of the continuous integral formulation.
Zero-IL Performance: The state-aware fee function successfully maintained zero IL for trades originating from $k_0$ . However, it charges significantly higher fees for large deviations compared to a constant 0.3% fee, effectively trading off trader costs for LP protection.
Slippage: Zero-IL fees induce slightly higher slippage for large trades, which is the necessary cost to fully compensate LPs.

5. Significance and Implications

Composability: By ensuring path independence, these fee structures restore deterministic outcomes for smart contracts interacting with the pool, a crucial requirement for complex DeFi composability.
Risk Modeling: LPs can now model returns based on aggregate volume rather than worrying about trade fragmentation strategies.
Protocol Design: The paper provides a blueprint for "Zero-IL" pools. While a universal solution is impossible, protocols can implement state-aware fees (e.g., via Uniswap V4 hooks) to target specific liquidity ranges, offering LPs perfect protection within those ranges.
Trade-offs: The work highlights a fundamental trade-off: eliminating IL requires dynamic, state-dependent fees that charge more for larger trades, potentially reducing liquidity depth for large traders.

In summary, this paper provides the first rigorous mathematical characterization of path-independent fees in CPMMs and demonstrates that while universal zero-IL is impossible, state-specific zero-IL is achievable through carefully designed, adaptive fee functions.